vla-mujoco-so101-push_cube_to_tray-v2.1
收藏Hugging Face2026-05-26 更新2026-05-26 收录
下载链接:
https://huggingface.co/datasets/CarlosMunoz0/vla-mujoco-so101-push_cube_to_tray-v2.1
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是使用LeRobot工具创建的机器人操作数据集,具体任务是在MuJoCo模拟环境中,使用so101机器人执行将立方体推送到托盘的操作。数据集包含100个训练episodes,总计75000帧,帧率为30fps。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。数据集包含以下特征:动作(6维浮点数组,表示机器人关节位置,包括肩部平移、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、末端执行器位置(x, y, z)和四元数(x, y, z, w));图像观测(包括RealSense彩色摄像头、RealSense深度摄像头、RealSense深度可视化摄像头和腕部摄像头的视频数据,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。
This dataset is created using the LeRobot tool for robot manipulation tasks, specifically involving pushing a cube to a tray with the so101 robot in the MuJoCo simulation environment. It contains 100 training episodes, totaling 75,000 frames at a frame rate of 30fps. The data is stored in parquet format, with a total data file size of 100MB and video file size of 200MB. The dataset includes the following features: actions (a 6-dimensional float array representing robot joint positions, including shoulder translation, shoulder lift, elbow bend, wrist bend, wrist rotation, and gripper position); state observations (a 13-dimensional float array including joint positions, end-effector position (x, y, z), and quaternion (x, y, z, w)); image observations (video data from RealSense RGB camera, RealSense depth camera, RealSense depth visualization camera, and wrist camera, all with a resolution of 480x640, 3 channels, encoded as av1); and metadata such as timestamp, frame index, episode index, index, and task index. All data is intended for training split only.
提供机构:
CarlosMunoz0创建时间:
2026-05-26
原始信息汇总
数据集概述
- 数据集名称: vla-mujoco-so101-push_cube_to_tray-v2.1
- 任务类别: 机器人 (robotics)
- 许可协议: Apache-2.0
- 创建工具: 使用 LeRobot 创建
数据集结构
| 属性 | 值 |
|---|---|
| 代码版本 | v3.0 |
| 机器人类型 | so101_mujoco |
| 总片段数 | 100 |
| 总帧数 | 75,000 |
| 总任务数 | 1 |
| 数据块大小 | 1,000 |
| 数据文件大小 | 100 MB |
| 视频文件大小 | 200 MB |
| 帧率 | 30 FPS |
| 数据分割 | 训练集 (0:100) |
数据特征
动作 (action)
- 数据类型: float32
- 维度: 6
- 名称: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
观测状态 (observation.state)
- 数据类型: float32
- 维度: 13
- 名称: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos, ee_pos_x, ee_pos_y, ee_pos_z, ee_quat_x, ee_quat_y, ee_quat_z, ee_quat_w
观测图像 (observation.images)
- 摄像头: realsense, realsense_depth, realsense_depth_vis, wrist_cam
- 分辨率: 480×640 像素
- 通道数: 3 (RGB)
- 视频编码: AV1
- 像素格式: yuv420p
- 帧率: 30 FPS
- 非深度图: 是
其他特征
- 时间戳 (
timestamp): float32, 维度1 - 帧索引 (
frame_index): int64, 维度1 - 片段索引 (
episode_index): int64, 维度1 - 索引 (
index): int64, 维度1 - 任务索引 (
task_index): int64, 维度1
引用信息
- BibTeX: [More Information Needed]



