grab_blue_cube_20260525_171034

该数据集是一个使用LeRobot创建的机器人任务数据集，专为训练设计。它包含20个episodes，总计17979帧，覆盖1个任务。数据特征丰富，包括动作数据（由6个浮点数表示，分别对应肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）、状态观测数据（同样为6个浮点数，与动作一致）、图像观测数据（来自固定摄像头和手眼摄像头的视频流，分辨率为480x640，3通道，帧率为30fps，编码为av1，格式为yuv420p），以及元数据如时间戳、帧索引、episode索引、索引和任务索引。数据集以parquet文件（总大小100MB）和视频文件（总大小200MB）形式存储，机器人类型为so_follower，适用于机器人任务学习和模型训练。

This dataset is a robotic task dataset developed using LeRobot, specifically designed for training purposes. It contains 20 episodes with a total of 17,979 frames, covering one single task. The dataset features rich multimodal data, including action data represented by 6 floating-point numbers corresponding to shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation, and gripper position respectively; state observation data, which also consists of 6 floating-point numbers consistent with the action data; image observation data sourced from video streams of fixed cameras and eye-in-hand cameras, with a resolution of 480×640, 3 channels, 30 frames per second (fps), encoded in AV1 format with the pixel format yuv420p; and metadata such as timestamps, frame index, episode index, sample index, and task index. The dataset is stored in Parquet files with a total size of 100 MB and video files with a total size of 200 MB. The robot type is so_follower, making it applicable for robotic task learning and model training.