Rick0331/openarm-tablewares-sort-merged3-pos

该数据集是一个机器人操作数据集，使用LeRobot创建，专门用于机器人技术研究。数据集记录了OpenArm Follower机器人在执行单一任务时的操作数据，包含178个完整 episodes，总计135,483帧，以30帧/秒的速率采集。数据格式包括Parquet数据文件和MP4视频文件。数据集特征丰富，包括：动作数据（8维，涵盖7个关节位置和夹爪位置）、观测状态（8维，同样包含关节和夹爪位置）、来自两个摄像头的视觉观测（前置摄像头和腕部摄像头，分辨率均为720x1280，RGB三通道）、时间戳、帧索引、episode索引、任务索引等。视频采用AV1编码，YUV420p像素格式。数据集适用于机器人控制、模仿学习、强化学习等任务。This dataset is a robotics manipulation dataset created using LeRobot, designed for robotics research. It records operational data from an OpenArm Follower robot performing a single task, comprising 178 complete episodes with a total of 135,483 frames captured at 30 frames per second. The data format includes Parquet data files and MP4 video files. The dataset features comprehensive data: action data (8-dimensional, covering 7 joint positions and gripper position), observation state (8-dimensional, also including joint and gripper positions), visual observations from two cameras (front camera and wrist camera, both with 720x1280 resolution, RGB three channels), timestamp, frame index, episode index, task index, etc. Videos are encoded in AV1 with YUV420p pixel format. The dataset is suitable for tasks such as robot control, imitation learning, and reinforcement learning.