YOLO2431/put_plate_to_rack_v4_full

该数据集使用LeRobot工具创建，专注于机器人学领域，特别是双手机器人（yam_bimanual）的操控任务。数据集包含301个episodes，总计286,658帧数据，对应903个视频文件，帧率为30fps。数据以parquet格式存储，仅包含训练集。数据集的特征丰富，包括状态观测（observation.state）和动作（action），均为32维浮点数向量，涵盖左右机械臂的3D位置坐标、6D旋转表示、夹持器状态以及各6个关节角度；视频观测来自三个摄像头视角：头部（head）、左手腕（left_wrist）和右手腕（right_wrist），每个视频为彩色（3通道），分辨率720x1280，使用h264编码。此外，数据集还包含时间戳、帧索引、episode索引等元数据。该数据集适用于机器人模仿学习、强化学习等研究，但主页和论文信息暂未提供。This dataset was created using the LeRobot tool and focuses on robotics, specifically bimanual robot (yam_bimanual) manipulation tasks. It contains 301 episodes, totaling 286,658 frames, corresponding to 903 video files with a frame rate of 30fps. The data is stored in parquet format and includes only a training split. The dataset features rich observations and actions: state observations (observation.state) and actions (action) are both 32-dimensional float32 vectors, covering 3D position coordinates, 6D rotation representations, gripper states, and 6 joint angles for both left and right arms; video observations come from three camera perspectives: head, left wrist, and right wrist, each video is color (3 channels) with a resolution of 720x1280, encoded using h264. Additionally, the dataset includes metadata such as timestamps, frame indices, and episode indices. It is suitable for research in robot imitation learning, reinforcement learning, etc., but homepage and paper information are currently not provided.