eval_shoulder_pan_10_minutes

该数据集是由LeRobot创建的机器人学习数据集，包含机器人动作、观察状态和图像数据等特征。数据集结构包括6个浮点型动作特征（肩部、肘部和手腕的位置）、6个浮点型状态观察特征、两个480x640 RGB视频图像（基础视角和左手腕视角），以及时间戳、帧索引、任务索引等元数据。数据以parquet格式存储，视频以mp4格式存储，帧率为10fps。This dataset is a robotic learning dataset created by LeRobot, which includes features such as robot actions, observation states, and image data. Its structure consists of 6 floating-point action features (positions of shoulder, elbow, and wrist joints), 6 floating-point state observation features, two 480x640 RGB video files captured from the base viewpoint and left wrist viewpoint, as well as metadata including timestamps, frame indices, task indices and other relevant information. The structured tabular data is stored in Parquet format, while the video files are saved in MP4 format with a frame rate of 10 FPS.