kunhsiang/eval_exp9_sc1var_grasp_the_big_box_20260528-183937
收藏资源简介:
该数据集使用LeRobot创建,专注于机器人技术领域,具体任务为抓取大盒子。数据集包含1个完整episode,总计2569帧,帧率为30fps。数据以Parquet文件格式存储,并包括视频文件。特征包括:动作(6个浮点数,对应机器人关节位置)、观察状态(6个浮点数,与动作相同)、多视角图像观察(前、上、夹爪三个视角,每个为480x640x3的视频,使用AV1编解码器)、时间戳、帧索引、episode索引等。数据集适用于机器人控制、模仿学习等研究。
This dataset was created using LeRobot and focuses on the robotics domain, specifically for the task of grasping the big box. It contains 1 complete episode with a total of 2569 frames at 30fps. The data is stored in Parquet format and includes video files. Features include: action (6 float values for robot joint positions), observation state (6 float values, same as action), multi-view image observations (front, top, and gripper views, each as 480x640x3 video with AV1 codec), timestamp, frame index, episode index, and others. The dataset is suitable for research in robot control, imitation learning, and related areas.




