Jiafei1224/eval_pi05_multi_packing_200k_03012026_snacks

该数据集是使用LeRobot创建的机器人数据集，专为机器人学习任务设计。数据集包含50个完整的episodes，总计189,142帧数据，涉及双机器人系统（bi_yam_follower）。数据特征包括：动作数据（14维浮点数组，表示左右机械臂的关节位置和夹爪位置）、状态观测（与动作相同的14维关节位置）、以及来自三个摄像头视角的视频观测（左、右、顶部，分辨率均为360x640，RGB三通道，帧率30fps）。此外，还包含时间戳、帧索引、episode索引等元数据。数据集以parquet格式存储，视频以mp4格式存储，适用于机器人控制、模仿学习等研究。This dataset was created using LeRobot and is designed for robotics learning tasks. It contains 50 full episodes with a total of 189,142 frames, involving a dual-robot system (bi_yam_follower). The features include: action data (14-dimensional float32 arrays representing joint positions and gripper positions for left and right arms), state observations (same 14-dimensional joint positions as actions), and video observations from three camera views (left, right, top, all with 360x640 resolution, 3 RGB channels, 30 fps). Additionally, metadata such as timestamp, frame index, and episode index are included. The dataset is stored in parquet format with videos in mp4, suitable for research in robot control, imitation learning, etc.