hjkso1406/multitask-so101-50eps-per-task-base

该数据集是一个用于机器人学习的多任务数据集，通过LeRobot工具创建。它包含150个集，总计113,268帧，帧率为30 fps，涉及3个不同任务。数据集结构包括训练分割（0-150集），数据以Parquet格式存储，总大小为100 MB，视频文件总大小为200 MB。特征包括机器人的动作（6个关节位置：肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）、观察状态（同样6个关节位置）、来自顶部和腕部摄像头的图像观察（分辨率480x640，3通道彩色视频，使用AV1编解码器），以及时间戳、帧索引、集索引、索引和任务索引等元数据。机器人类型为so_follower，代码库版本为v3.0。数据集适用于机器人控制、视觉感知和多任务学习研究。

This dataset is a multitask dataset for robotics learning, created using the LeRobot tool. It contains 150 episodes, totaling 113,268 frames at 30 fps, involving 3 different tasks. The dataset structure includes a training split (episodes 0-150), with data stored in Parquet format, total data size of 100 MB, and video files totaling 200 MB. Features include robot actions (6 joint positions: shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper position), observation states (same 6 joint positions), image observations from top and wrist cameras (resolution 480x640, 3-channel color video, using AV1 codec), and metadata such as timestamp, frame index, episode index, index, and task index. The robot type is so_follower, and the codebase version is v3.0. The dataset is suitable for research in robot control, visual perception, and multitask learning.