stevenworkspace/eval_take_4
收藏Hugging Face2026-05-28 更新2026-05-31 收录
下载链接:
https://hf-mirror.com/datasets/stevenworkspace/eval_take_4
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个机器人学数据集,使用LeRobot创建,专门针对mobileai_robot机器人类型。数据集包含1个episode、689帧数据和1个任务,以30帧每秒(fps)的帧率记录。数据特征包括动作(action)和观察状态(observation.state),其中动作和状态均由16个浮点数值组成,代表左右机械臂的关节位置(如left_joint_0.pos、right_joint_0.pos等)和速度参数(x.vel、theta.vel)。观察部分还包括三个摄像头的视频数据:cam_high、cam_left_wrist和cam_right_wrist,每个视频的分辨率为480x640像素,3个颜色通道,使用AV1编码和yuv420p像素格式,不包含音频。此外,数据集包含时间戳、帧索引、episode索引、索引和任务索引等元数据,用于数据组织和跟踪。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,适用于机器人控制、计算机视觉和机器学习任务。
This dataset is a robotics dataset created using LeRobot, specifically for the mobileai_robot type. It contains 1 episode, 689 frames, and 1 task, recorded at a frame rate of 30 frames per second (fps). The data features include action and observation.state, both consisting of 16 floating-point values representing joint positions (e.g., left_joint_0.pos, right_joint_0.pos) and velocity parameters (x.vel, theta.vel) for the left and right robotic arms. The observation part also includes video data from three cameras: cam_high, cam_left_wrist, and cam_right_wrist, each with a resolution of 480x640 pixels, 3 color channels, encoded in AV1 with yuv420p pixel format, and without audio. Additionally, the dataset contains metadata such as timestamp, frame index, episode index, index, and task index for data organization and tracking. The data is stored in parquet format, with a total data file size of 100MB and video file size of 200MB, suitable for robotics control, computer vision, and machine learning tasks.
提供机构:
stevenworkspace


