lmerkle/multi_task_20260529_221601
收藏资源简介:
该数据集是使用LeRobot创建的机器人数据集,专注于多任务机器人操作。数据集包含机器人的动作和观测数据,具体特征包括动作(如肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置)、观测状态(与动作相同的关节位置)、前摄像头图像(480x640分辨率,RGB视频)和腕部摄像头图像(相同规格)。数据以parquet文件格式存储,视频文件为mp4格式,帧率为30fps。机器人类型为so_follower,数据集总共有1个任务、1个剧集和900帧,数据总大小约为300MB(100MB数据文件和200MB视频文件)。该数据集适用于机器人学习和控制研究,特别是基于视觉的机器人操作任务。
This dataset was created using LeRobot and focuses on multi-task robot manipulation. It includes robot action and observation data, with features such as action (e.g., shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper positions), observation state (same joint positions as action), front camera images (480x640 resolution, RGB video), and wrist camera images (same specifications). Data is stored in parquet files, with video files in mp4 format at 30fps. The robot type is so_follower, and the dataset contains 1 task, 1 episode, and 900 frames in total, with an approximate size of 300MB (100MB for data files and 200MB for video files). This dataset is suitable for robot learning and control research, particularly for vision-based robot manipulation tasks.




