amber2713/merged-record-test_4

该数据集使用LeRobot创建，专注于机器人技术领域，具体涉及so101_follower机器人类型。数据集包含192个episodes和70002帧数据，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。数据特征包括动作（如肩部、肘部、腕部和夹爪的位置）、观察状态（机器人状态和来自前视、腕部和上方摄像头的图像，分辨率为480x640，3通道RGB视频），以及时间戳、帧索引、episode索引等元数据。数据集仅包含训练分割（0:192），适用于机器人控制、视觉感知和强化学习等任务。

This dataset was created using LeRobot and focuses on the robotics domain, specifically involving the so101_follower robot type. It contains 192 episodes and 70,002 frames of data, stored in parquet format, with a total data file size of 100MB and video file size of 200MB, at a frame rate of 30fps. The features include actions (such as positions of shoulder, elbow, wrist, and gripper), observation states (robot state and images from front, wrist, and up cameras, with a resolution of 480x640 and 3-channel RGB video), as well as metadata like timestamps, frame indices, and episode indices. The dataset only includes a training split (0:192) and is suitable for tasks such as robot control, visual perception, and reinforcement learning.