jjr1007/may7_merged

该数据集是使用LeRobot创建的机器人数据集，专门用于机器人技术任务。数据集包含255个episodes，总计234121帧，以30帧/秒的速率采集。数据以parquet文件和MP4视频格式存储，总数据文件大小为100MB，视频文件大小为200MB。数据集的特征包括：6维动作空间（对应机械臂的肩部旋转、肩部抬升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）、6维状态观测（与动作空间相同）、来自前摄像头的RGB视频观测（分辨率为1080x1920，3个通道），以及其他元数据如时间戳、帧索引、episode索引和任务索引。机器人类型为so_follower，表明数据集可能用于机械臂的跟随或控制任务。数据集仅包含训练分割（所有255个episodes），适用于机器人学习和模仿学习应用。

This dataset is a robotics dataset created using LeRobot, designed for robotics tasks. It contains 255 episodes with a total of 234121 frames, captured at 30 frames per second. The data is stored in parquet files and MP4 videos, with a total data file size of 100MB and video file size of 200MB. The dataset features include: a 6-dimensional action space (corresponding to robot arm joint positions for shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper), a 6-dimensional state observation (same as the action space), RGB video observations from a front camera (resolution 1080x1920 with 3 channels), and other metadata such as timestamp, frame index, episode index, and task index. The robot type is so_follower, suggesting the dataset may be used for follower or control tasks with a robotic arm. The dataset includes only a training split (all 255 episodes) and is suitable for robotics learning and imitation learning applications.