YOLO2431/clean_up_table_v2_final

该数据集是一个用于机器人研究的双手机器人（Yam Bimanual）数据集，由LeRobot工具创建。它包含300个episodes，总计246,897帧数据，覆盖4个不同任务，并提供了900个视频文件。数据以30fps的帧率采集，仅包含训练分割（0:300）。数据集的结构化特征包括：状态观测（observation.state），维度为32，表示左右机械臂的位置、旋转（使用6D表示）、夹爪状态和关节角度；动作（action），维度为32，与状态观测对应；图像观测来自三个摄像头视角（头部、左腕、右腕），每个视频分辨率为360x640，3通道RGB格式；此外，还包含时间戳、帧索引、episode索引、全局索引和任务索引等元数据。该数据集适用于强化学习、模仿学习、机器人控制算法开发和计算机视觉任务的研究。This dataset is a bimanual robot (Yam Bimanual) dataset for robotics research, created using the LeRobot tool. It contains 300 episodes, totaling 246,897 frames, covering 4 distinct tasks, and includes 900 video files. The data is collected at a frame rate of 30fps, with only a training split (0:300). The structured features of the dataset include: state observations (observation.state) with a dimension of 32, representing the position, rotation (using 6D representation), gripper status, and joint angles of the left and right robotic arms; actions (action) with a dimension of 32, corresponding to the state observations; image observations from three camera perspectives (head, left wrist, right wrist), each video with a resolution of 360x640 and 3-channel RGB format; additionally, it includes metadata such as timestamps, frame indices, episode indices, global indices, and task indices. This dataset is suitable for research in reinforcement learning, imitation learning, robot control algorithm development, and computer vision tasks.