hjkso1406/eval_xvla-multitask-so101-70eps-per-task-base-aug_pick-place
收藏资源简介:
该数据集使用LeRobot创建,是一个机器人任务数据集,专注于多任务操作。数据集包含10个episodes,总计10442帧,帧率为30fps,数据以parquet文件格式存储,视频以mp4格式存储。特征包括动作(6个关节位置:肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转、夹爪位置)、观察状态(6个关节位置)、两个图像观察(分辨率为480x640的RGB视频,用于视觉输入),以及时间戳、帧索引、episode索引、任务索引等元数据。数据集适用于机器人控制、强化学习或模仿学习任务,特别是涉及视觉和动作协调的场景。
This dataset was created using LeRobot and is a robotics task dataset focused on multitask operations. It contains 10 episodes, totaling 10442 frames with a frame rate of 30fps. The data is stored in parquet files, and videos are in mp4 format. Features include actions (6 joint positions: shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, gripper position), observation states (6 joint positions), two image observations (RGB videos with resolution 480x640 for visual input), and metadata such as timestamp, frame index, episode index, and task index. The dataset is suitable for robotics control, reinforcement learning, or imitation learning tasks, particularly those involving visual and action coordination.




