record-test_20260703_155639

该数据集是使用LeRobot创建的，专为机器人学任务设计。它包含机器人的动作和观测数据：动作和观测状态均为6维浮点数组，表示机器人关节位置，具体包括肩部平移、肩部抬升、肘部弯曲、手腕弯曲、手腕旋转和夹爪位置；观测图像来自两个摄像头（cam_0和cam_1），每帧为480x640x3的视频数据，帧率为30fps，使用AV1编码。元数据包括时间戳、帧索引、回合索引、索引和任务索引。数据集总共有5个回合、3780帧和1个任务，文件格式为Parquet，总数据大小为100MB，视频文件大小为200MB。适用于训练和评估机器人控制模型，机器人类型为so_follower。

This dataset is created using LeRobot and is primarily used for robotics tasks. It includes robot action and observation data: both action and observation.state are 6-dimensional float arrays representing robot joint positions, such as shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position; observation images come from two cameras (cam_0 and cam_1), each frame being 480x640x3 video data at 30fps with AV1 encoding. Other metadata includes timestamp, frame index, episode index, index, and task index. The dataset consists of 5 episodes, 3780 frames, and 1 task, in Parquet format, with a total data size of 100MB and video file size of 200MB. It is suitable for training and evaluating robot control models, with the robot type being so_follower.