mujoco-so101-take_out_box-mouse-v1.0
收藏Hugging Face2026-05-18 更新2026-05-18 收录
下载链接:
https://huggingface.co/datasets/Juanchix/mujoco-so101-take_out_box-mouse-v1.0
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个机器人数据集,使用LeRobot工具创建,专门用于机器人学习任务。它基于Mujoco模拟环境,机器人类型为so101_mujoco,包含70个episodes,总计52496帧,涉及1个任务。数据以parquet格式存储,总数据文件大小为100MB,视频文件大小为200MB,帧率为30fps。数据集包含多个特征:动作数据(6维浮点数组,表示机器人关节位置,如肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置)、状态观测(13维浮点数组,包括关节位置和末端执行器位置与四元数)、多种图像观测(如realsense、realsense_depth、realsense_depth_vis和wrist_cam视频,分辨率均为480x640,3通道,编码为av1),以及时间戳、帧索引、episode索引、索引和任务索引等元数据。所有数据仅用于训练分割(train split: 0:70)。
This dataset is a robotics dataset created using the LeRobot tool, specifically designed for robot learning tasks. It is based on the Mujoco simulation environment, with a robot type of so101_mujoco, containing 70 episodes, totaling 52496 frames, and involving 1 task. The data is stored in parquet format, with a total data file size of 100MB, a video file size of 200MB, and a frame rate of 30fps. The dataset includes multiple features: action data (6-dimensional float arrays representing robot joint positions, such as shoulder translation, shoulder lift, elbow bend, wrist bend, wrist rotation, and gripper position), state observations (13-dimensional float arrays including joint positions and end-effector positions with quaternions), various image observations (such as realsense, realsense_depth, realsense_depth_vis, and wrist_cam videos, all with a resolution of 480x640, 3 channels, encoded in av1), and metadata such as timestamps, frame indices, episode indices, indices, and task indices. All data is exclusively for the training split (train split: 0:70).
提供机构:
Juanchix
创建时间:
2026-05-18
原始信息汇总
数据集概述
该数据集是一个基于 MuJoCo 仿真环境的机器人操控数据集,主要针对 so101_mujoco 机器人执行“取盒子”任务。
基本信息
- 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 创建工具: LeRobot
- 机器人类型: so101_mujoco
- 总片段数: 80
- 总帧数: 59996
- 总任务数: 1
- 数据集拆分: 训练集 (0:80)
- 帧率 (FPS): 30
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
数据特征
数据特征包含动作、观察状态、多视角图像、时间戳及索引信息,具体如下:
- action (动作):
- 类型: float32
- 维度: 6
- 名称:
shoulder_pan.pos,shoulder_lift.pos,elbow_flex.pos,wrist_flex.pos,wrist_roll.pos,gripper.pos
- observation.state (观察状态):
- 类型: float32
- 维度: 13
- 名称: 包含6个关节位置 (
shoulder_pan.pos等) 及7个末端执行器位姿信息 (ee_pos_x,ee_pos_y,ee_pos_z,ee_quat_x,ee_quat_y,ee_quat_z,ee_quat_w)
- observation.images (观察图像):
- realsense: 彩色视频, 分辨率 480x640, 3通道, 编码 av1, 帧率 30
- realsense_depth: 深度视频, 分辨率 480x640, 3通道, 编码 av1, 帧率 30
- realsense_depth_vis: 深度可视化视频, 分辨率 480x640, 3通道, 编码 av1, 帧率 30
- wrist_cam: 腕部摄像头视频, 分辨率 480x640, 3通道, 编码 av1, 帧率 30
- 其他特征:
timestamp(时间戳),frame_index(帧索引),episode_index(片段索引),index(全局索引),task_index(任务索引)
数据结构
- 代码库版本: v3.0
- 数据路径格式:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径格式:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4



