nanyong/hd_cube_data2

该数据集是一个机器人学数据集，使用LeRobot创建，包含so101_follower机器人的模仿学习或控制数据。数据集共有261个训练片段，总计164,949帧，帧率为30fps。数据特征包括：6维动作和状态向量（对应肩部平移、肩部抬升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置），以及来自腕部和前部摄像头的视频观测（分辨率480x640，3通道，AV1编码）。此外，还包含时间戳、帧索引、片段索引和任务索引等元数据。数据集结构以Parquet文件存储，总数据文件大小为100MB，视频文件大小为200MB，适用于机器人行为克隆或强化学习任务。This dataset is a robotics dataset created using LeRobot, containing imitation learning or control data for the so101_follower robot. It consists of 261 training episodes with a total of 164,949 frames at 30fps. Features include: 6-dimensional action and state vectors (corresponding to shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper positions), video observations from wrist and front cameras (resolution 480x640, 3 channels, AV1 codec). Additionally, metadata such as timestamp, frame index, episode index, and task index are included. The dataset is stored in Parquet files, with a total data file size of 100MB and video file size of 200MB, suitable for robot behavior cloning or reinforcement learning tasks.