stevenworkspace/eval_take_13
收藏资源简介:
该数据集是一个用于机器人学习和控制的数据集,基于LeRobot工具创建。数据集包含来自mobileai_robot机器人的多模态传感器数据,总共有1个剧集、1040帧和1个任务。数据以parquet文件格式存储,并附带视频文件。特征包括动作数据(16维,涵盖左右关节位置和速度)、观察状态数据(16维,与动作数据相同)、以及来自三个摄像头的图像数据(cam_high、cam_left_wrist、cam_right_wrist),图像分辨率为480x640,3通道,帧率30fps。此外,还包含时间戳、帧索引、剧集索引等元数据。数据集适用于机器人模拟、控制算法训练和视觉-动作学习等任务。
This dataset is designed for robot learning and control, created using the LeRobot tool. It includes multimodal sensor data from a mobileai_robot, with a total of 1 episode, 1040 frames, and 1 task. The data is stored in parquet files and accompanied by video files. Features encompass action data (16 dimensions, covering left and right joint positions and velocities), observation state data (16 dimensions, identical to action data), and image data from three cameras (cam_high, cam_left_wrist, cam_right_wrist) with a resolution of 480x640, 3 channels, and a frame rate of 30fps. Additionally, metadata such as timestamps, frame indices, and episode indices are included. The dataset is suitable for tasks like robot simulation, control algorithm training, and vision-action learning.




