rollout_recap_stackblocks_iter0_20260704_173438

该数据集是由LeRobot创建的机器人控制数据集，专用于机器人学习任务。它包含2个剧集、4351帧和1个任务，数据以parquet格式存储。特征包括动作（一个6维浮点数组，对应机器人关节位置，具体为肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）、观察状态（与动作相同的6维浮点数组）、观察图像（从顶部和腕部视角采集的视频，分辨率为480x640，3通道，帧率30fps，使用av1编码）、干预（布尔值）、下一帧成功标志（布尔值）、时间戳（浮点数）、帧索引（整数）、剧集索引（整数）、索引（整数）和任务索引（整数）。机器人类型为so_follower，数据集被分割为训练集（范围0:2）。数据文件总大小为100MB，视频文件总大小为200MB。

This dataset is a robot control dataset created by LeRobot for robot learning tasks. It contains 2 episodes, 4351 frames, and 1 task, with data stored in parquet format. Features include action (a 6-dimensional float array corresponding to robot joint positions such as shoulder translation, shoulder lift, elbow bend, wrist bend, wrist rotation, and gripper position), observation state (the same 6-dimensional float array as action), observation images (videos from top and wrist perspectives with a resolution of 480x640, 3 channels, 30fps frame rate, encoded using av1), intervention (boolean), next frame success flag (boolean), timestamp (float), frame index (integer), episode index (integer), index (integer), and task index (integer). The robot type is so_follower, and the dataset is split into training set (0:2). The total data file size is 100MB, and the total video file size is 200MB.