ankile/real01c-insert-marker-d1-blind-eval-r1-25k-sobol20
收藏资源简介:
该数据集是使用LeRobot创建的机器人数据集,专为机器人学习和控制任务设计。数据集基于Franka机器人,包含80个完整的情节(episodes),总计19296帧数据,帧率为15fps,仅包含一个任务。数据以Parquet格式存储,并附带视频文件。数据集特征丰富,包括观测状态(如笛卡尔位置、关节位置、速度、夹爪位置)、动作(如笛卡尔速度、关节速度、夹爪动作)、奖励、完成标志、步骤剩余数、策略ID、回合ID等。此外,还提供两个摄像头(ID: 18650758_left和25916956_left)的视频流,分辨率为480x640,编码为h264,用于视觉观测。数据集适用于机器人强化学习、模仿学习等研究,支持训练和评估机器人控制策略。
This dataset is a robotics dataset created using LeRobot, designed for robot learning and control tasks. It is based on the Franka robot and includes 80 complete episodes, totaling 19296 frames with a frame rate of 15fps, and contains only one task. The data is stored in Parquet format and accompanied by video files. The dataset features a wide range of attributes, including observation states (such as Cartesian position, joint position, velocity, gripper position), actions (such as Cartesian velocity, joint velocity, gripper action), reward, done flag, steps to go, policy ID, round ID, and more. Additionally, it provides video streams from two cameras (IDs: 18650758_left and 25916956_left) with a resolution of 480x640, encoded in h264, for visual observations. The dataset is suitable for research in robotics reinforcement learning, imitation learning, and other areas, supporting the training and evaluation of robot control policies.




