DAVIAN-Robotics/peg_in_hole

该数据集是一个机器人操作数据集，使用LeRobot工具创建。数据集基于Franka机器人，包含100个episodes，总计24190帧，覆盖5个不同任务。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20fps。数据集的特征包括：状态观测（29维浮点数组，包含关节位置、速度、末端执行器位姿、抓取状态等）、动作（8维浮点数组，表示末端执行器位置、姿态和夹爪控制）、来自三个摄像头（手腕、左后、右后）的图像观测（视频格式，分辨率480x640，3通道RGB），以及时间戳、帧索引、episode索引、任务索引等元数据。数据集仅包含训练集，用于机器人学习和控制任务。This dataset is a robotic manipulation dataset created using the LeRobot tool. It is based on a Franka robot and includes 100 episodes, totaling 24190 frames, covering 5 distinct tasks. The data is stored in parquet format, with a total data file size of 100MB and video file size of 200MB, at a frame rate of 20fps. The dataset features include: state observation (a 29-dimensional float32 array containing joint positions, velocities, end-effector pose, grasp state, etc.), action (an 8-dimensional float32 array representing end-effector position, orientation, and gripper control), image observations from three cameras (wrist, left_back, right_back) in video format (resolution 480x640, 3-channel RGB), and metadata such as timestamp, frame index, episode index, and task index. The dataset only contains a training split and is intended for robot learning and control tasks.