rollout_perfect_50episodes_2cm_20260612_160334

这是一个使用LeRobot创建的机器人数据集，基于so_follower机器人类型。数据集包含5个总集数和2532个总帧数。特征包括动作（由6个关节位置组成：肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转、夹爪位置）、观测状态（与动作相同的6个关节位置），以及两个摄像头图像：顶部摄像头和手腕摄像头，均为480x640分辨率、3通道彩色视频，帧率30fps，使用AV1编码。此外，数据集还包含干预信号、时间戳、帧索引、集索引、任务索引等元数据。数据以Parquet格式存储，视频文件为MP4格式，总数据大小约100MB，视频大小约200MB。该数据集专为训练和评估机器人控制任务设计，支持多模态输入，结合关节状态和视觉图像。This is a robot dataset created using LeRobot, based on the so_follower robot type. The dataset consists of 5 total episodes and 2532 total frames. Features include actions (comprising 6 joint positions: shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position), observation states (the same 6 joint positions), and two camera images: a top camera and a wrist camera, both with 480x640 resolution, 3-channel color video, 30fps frame rate, encoded using AV1. Additionally, the dataset includes metadata such as intervention signals, timestamps, frame indices, episode indices, and task indices. Data is stored in Parquet format, with video files in MP4 format; the total data size is approximately 100MB, and the video size is about 200MB. This dataset is designed for training and evaluating robot control tasks, supporting multimodal input that combines joint states and visual images.