SortPaper_20260513_162720

该数据集由LeRobot创建，专用于机器人学任务，特别是分拣操作（SortPaper）。它包含一个总时长为790帧的单个episode，帧率为30fps。数据集特征包括动作（action）和观测（observation），其中动作和状态观测均包含6个浮点数值，分别对应机械臂的关节位置：肩部平移、肩部抬升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置。观测部分还包括来自前部和顶部摄像头的图像数据，分辨率为480x640，3通道，视频编码为av1。此外，数据集提供时间戳、帧索引、episode索引等元数据。机器人类型为so_follower，数据以parquet格式存储，总数据文件大小为100MB，视频文件大小为200MB。该数据集适用于机器人控制、模仿学习或强化学习等任务。

This dataset was developed by LeRobot, tailored for robotics tasks, with a particular focus on the SortPaper sorting operation. It comprises a single episode with a total of 790 frames, recorded at 30 frames per second (fps). The dataset includes action and observation data: both action signals and state observations consist of 6 floating-point values, which correspond to the robotic arm's joint positions in sequence: shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position. The observation data also incorporates image data from the front and top cameras, with a resolution of 480×640, 3 color channels, and encoded using the AV1 video format. Additionally, the dataset provides metadata such as timestamps, frame indices, and episode indices. The robotic platform used is so_follower, and the data is stored in Parquet format, with a total data file size of 100 MB and a separate video file size of 200 MB. This dataset is applicable to tasks including robot control, imitation learning, and reinforcement learning.