CoRL2026-CSI/SO101_teleop_close_drawer_100epi_wristfix_50epi_8to2

该数据集是一个机器人学数据集，使用LeRobot创建，专为机器人控制任务设计。数据集包含50个情节，总计7867帧，帧率为30 fps，涉及一个任务（推测为关闭抽屉）。数据以parquet格式存储，总数据文件大小为100 MB，视频文件大小为200 MB。特征包括动作（6个浮点数，表示肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）、观测状态（与动作相同的6个浮点数）、图像观测（左腕和顶视图的两个视频，分辨率均为480x640，3通道，编码为h264，无音频）以及时间戳、帧索引、情节索引等元数据。数据集用于训练和评估机器人模型，支持从视觉和状态信息中学习控制策略。This dataset is a robotics dataset created using LeRobot, designed for robot control tasks. It contains 50 episodes with a total of 7867 frames at 30 fps, involving one task (likely closing a drawer). The data is stored in parquet format, with a total data file size of 100 MB and video file size of 200 MB. Features include actions (6 float values representing shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper positions), observation states (same 6 float values as actions), image observations (two videos from left wrist and topview, both with 480x640 resolution, 3 channels, encoded as h264, without audio), and metadata such as timestamp, frame index, episode index, etc. The dataset is intended for training and evaluating robot models, supporting learning control policies from visual and state information.