ki_left_task_22_bi_1

该数据集使用LeRobot创建，专注于机器人领域，特别是双手机器人（bi_so_follower）的控制与感知。它包含30个episodes，总计22416帧，涉及单一任务，数据以parquet格式存储，总数据文件大小为100MB，视频文件大小为200MB，帧率为30fps。数据集的特征包括：动作数据（12个浮点值，表示左右机械臂的关节位置，如肩部、肘部、腕部和夹爪）、观察状态（与动作相同的12个关节位置）、来自三个摄像头的图像观测（左前、左腕和右腕，分辨率均为360x640，3通道RGB视频，使用h264编码），以及时间戳、帧索引、episode索引等元数据。所有数据仅用于训练分割，适用于机器人学习、模仿学习或强化学习任务。

This dataset is created using LeRobot and focuses on the robotics domain, specifically the control and perception of a dual-arm robot (bi_so_follower). It contains 30 episodes, totaling 22416 frames, involving a single task, with data stored in parquet format. The total data file size is 100MB, video file size is 200MB, and frame rate is 30fps. The features of the dataset include: action data (12 floating-point values representing joint positions of the left and right robotic arms, such as shoulder, elbow, wrist, and gripper), observation state (the same 12 joint positions as the actions), image observations from three cameras (left front, left wrist, and right wrist, all with a resolution of 360x640, 3-channel RGB video, encoded with h264), and metadata such as timestamps, frame indices, and episode indices. All data is only for training split, suitable for robot learning, imitation learning, or reinforcement learning tasks.