simdata2

该数据集是使用LeRobot框架创建的，专注于机器人学中的螺母组装方形任务（nutassemblysquare）。它包含panda机器人的演示数据，适用于行为克隆（如BC-RNN）等任务。数据集结构包括3个episodes、600帧和1个任务，数据以parquet文件格式存储，视频以mp4格式存储。特征包括机器人状态观察（9个浮点数，表示关节和夹爪位置）、图像观察（来自agentview和wrist视角的480x640 RGB视频，帧率为30fps）、动作（7个浮点数，控制关节和夹爪）、以及时间戳、帧索引、episode索引等元数据。数据集总大小为100MB（数据文件）和200MB（视频文件），主要用于机器人模仿学习研究。This dataset is created using the LeRobot framework and focuses on the nut assembly square task in robotics. It contains demonstration data from a panda robot, suitable for tasks such as behavior cloning (e.g., BC-RNN). The dataset structure includes 3 episodes, 600 frames, and 1 task, with data stored in parquet format and videos in mp4 format. Features include robot state observations (9 floating-point numbers representing joint and gripper positions), image observations (480x640 RGB videos from agentview and wrist perspectives at 30fps), actions (7 floating-point numbers controlling joints and gripper), and metadata such as timestamps, frame indices, and episode indices. The total dataset size is 100MB for data files and 200MB for video files, intended for robot imitation learning research.