KS325/open-upper-drawer-r2-emp-train-expand

该数据集是一个机器人控制数据集，使用LeRobot工具创建。数据集包含28个训练episodes，总计20503帧，视频帧率为30fps。数据特征包括机器人的6维动作（肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转、夹爪位置）和状态观测（与动作相同的6维关节位置），以及两个摄像头（camera1和camera2）的视频流观测（分辨率480x640，RGB三通道）。此外，还包含时间戳、帧索引、episode索引、任务索引等元数据。数据集适用于机器人学习任务，如模仿学习或强化学习，用于训练机器人执行特定操作（如打开抽屉）。数据以parquet格式存储，视频以mp4格式存储，遵循Apache 2.0许可证。

This dataset is a robotics control dataset created using the LeRobot tool. It contains 28 training episodes with a total of 20,503 frames and a video frame rate of 30 fps. The features include 6-dimensional robot actions (shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, gripper positions) and state observations (same 6-dimensional joint positions as actions), along with video stream observations from two cameras (camera1 and camera2, resolution 480x640, RGB three channels). Additionally, it includes metadata such as timestamps, frame indices, episode indices, and task indices. The dataset is suitable for robotics learning tasks, such as imitation learning or reinforcement learning, to train robots for specific operations (e.g., opening drawers). Data is stored in parquet format, videos in mp4 format, and it is licensed under Apache 2.0.