tingtu0721/record-dataset-300-2cameras

该数据集使用LeRobot工具创建，是一个机器人控制数据集，专门用于so_follower机器人类型。数据集包含300个训练episodes，总计170,862帧数据，帧率为30fps。数据以parquet文件格式存储，总数据文件大小为100MB，视频文件大小为200MB。数据集结构包括动作数据（包含6个关节的位置：shoulder_pan.pos、shoulder_lift.pos、elbow_flex.pos、wrist_flex.pos、wrist_roll.pos、gripper.pos）、观测状态数据（与动作数据相同的关节位置）、来自两个摄像头（front和top）的视频观测（分辨率为480x640，编码为av1），以及时间戳、帧索引、episode索引等元数据。数据集仅包含训练分割，用于机器人学习和控制任务。This dataset was created using the LeRobot tool and is a robotics control dataset specifically for the so_follower robot type. It contains 300 training episodes, totaling 170,862 frames at 30fps. The data is stored in parquet format, with a total data file size of 100MB and video file size of 200MB. The dataset structure includes action data (with positions of 6 joints: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos), observation state data (same joint positions as actions), video observations from two cameras (front and top) with a resolution of 480x640 and av1 encoding, and metadata such as timestamp, frame index, and episode index. The dataset includes only a training split and is intended for robot learning and control tasks.