asdl-unist/eval_pi0-zeroshot-0

该数据集是一个机器人演示数据集，使用LeRobot创建。它包含一个完整的训练集，涉及so_follower机器人类型。数据集共有1个任务、1个完整片段和1461帧数据，数据以Parquet文件格式存储，总数据文件大小为100MB，视频文件大小为200MB，帧率为30fps。数据特征包括机器人的6维动作（如肩部平移、肩部提升、肘部弯曲等位置控制）和观测状态（同样为6维关节位置）。观测图像包括顶部摄像头和手腕摄像头拍摄的RGB视频，分辨率均为480x640，3通道，视频编码为av1。此外，数据集还包含时间戳、帧索引、片段索引、任务索引等元数据。该数据集适用于机器人学习任务，如模仿学习或强化学习。

This dataset is a robotics demonstration dataset created using LeRobot. It contains a complete training set involving the so_follower robot type. The dataset comprises 1 task, 1 full episode, and 1461 frames of data, stored in Parquet format. The total data file size is 100MB, and the video file size is 200MB, with a frame rate of 30fps. Data features include 6-dimensional robot actions (e.g., positions for shoulder pan, shoulder lift, elbow flex, etc.) and observation states (also 6-dimensional joint positions). Observation images include RGB videos from top and wrist cameras, both with a resolution of 480x640 and 3 channels, encoded in av1. Additionally, the dataset includes metadata such as timestamps, frame indices, episode indices, and task indices. This dataset is suitable for robotics learning tasks, such as imitation learning or reinforcement learning.