palanivelg123/pick_place_v3

该数据集是一个用于机器人任务（具体为拾放操作）的模仿学习数据集，使用LeRobot平台创建。数据集包含50个训练片段，总计26777帧数据，以30fps的帧率录制。数据以parquet文件格式存储，视频以mp4格式存储。数据集记录了机器人（so_follower类型）在执行单一任务时的状态和动作信息。具体特征包括：动作（action）——包含6个关节（肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转、夹爪）的位置信息；观测状态（observation.state）——与动作相同的6个关节位置；图像观测——来自顶部摄像头（observation.images.top）和手腕摄像头（observation.images.wrist）的RGB视频，分辨率均为480x640；以及时间戳、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控制、行为克隆或强化学习等研究。

This dataset is an imitation learning dataset for robotics tasks (specifically pick-and-place operations), created using the LeRobot platform. It contains 50 training episodes with a total of 26,777 frames, recorded at 30 fps. Data is stored in parquet format, and videos are stored in mp4 format. The dataset records state and action information of a robot (type: so_follower) performing a single task. Key features include: action—positions of 6 joints (shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, gripper); observation.state—the same 6 joint positions as in action; image observations—RGB videos from a top camera (observation.images.top) and a wrist camera (observation.images.wrist), both with a resolution of 480x640; and metadata such as timestamp, frame index, episode index, and task index. This dataset is suitable for research in robot control, behavior cloning, or reinforcement learning.