so101_desk_trash2

该数据集由LeRobot创建，专门用于机器人学习任务。它包含8个episodes，总计3461帧，帧率为30fps。数据以parquet文件和mp4视频文件形式存储，总大小为300MB（其中数据文件100MB，视频文件200MB）。数据集特征包括：动作（6维浮点数组，对应机械臂的关节位置：shoulder_pan.pos、shoulder_lift.pos、elbow_flex.pos、wrist_flex.pos、wrist_roll.pos、gripper.pos）、状态观测（与动作相同的6维关节位置）、前摄像头图像（480x640 RGB视频，无音频），以及时间戳、帧索引、episode索引、索引和任务索引。数据集仅包含一个任务类型，机器人类型为“so_follower”。数据分割为训练集，涵盖episodes 0到7。

This dataset was created by LeRobot for robot learning tasks. It contains 8 episodes with a total of 3461 frames at a frame rate of 30fps. The data is stored in parquet files and mp4 video files, with a total size of 300MB (100MB for data files and 200MB for video files). Features include actions (6-dimensional floating-point arrays corresponding to the joint positions of the robotic arm: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos), state observations (the same 6-dimensional joint positions as actions), front camera images (480x640 RGB video without audio), as well as timestamps, frame indices, episode indices, indices, and task indices. The dataset includes only one task type, with a robot type of so_follower. The data is split into a training set covering episodes 0 to 7.