stack_3_box_30x3030
收藏官方服务:
资源简介:
该数据集由LeRobot创建,用于机器人操作任务,具体为使用SO-101领导遥操作在MuJoCo模拟环境中堆叠3个盒子(尺寸30x30x30)。数据集包含来自一个episode的1116帧,动作和观察状态由6个关节位置(肩部、肘部、腕部、夹爪)组成,以及两个摄像头(顶部和正面)的视频图像,分辨率为480x640。机器人类型为so_follower。
This dataset is created by LeRobot for a robot manipulation task, specifically stacking 3 boxes (size 30x30x30) in a MuJoCo simulation environment using SO-101 leader teleoperation. The dataset contains 1116 frames from one episode, with actions and observation states consisting of 6 joint positions (shoulder, elbow, wrist, gripper) and video images from two cameras (top and front) at a resolution of 480x640. The robot type is so_follower.
提供机构:
anhquoc29创建时间:
2026-09-09
原始信息汇总
- 名称: stack_3_box_30x3030
- 许可证: apache-2.0
- 任务类别: 机器人技术(robotics)
- 标签: LeRobot, sim, mujoco, so101, leader-teleop, stack-3-blocks, 30x30x30, degree, 640x480
- 创建工具: 由LeRobot(https://github.com/huggingface/lerobot)创建
数据集结构
- 版本: codebase_version v3.0
- 帧率: 30 fps
- 总片段数: 1
- 总帧数: 1116
- 总任务数: 1
- 数据文件大小: 100 MB(parquet格式)
- 视频文件大小: 200 MB(mp4格式)
- 数据路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4 - 机器人类型: so_follower
- 数据划分: train(0:1,全部数据)
特征信息
- observation.state(
float32, 形状 [6]):- shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
- observation.images.top(视频, 480x640x3):
- 编码格式av1,像素格式yuv420p,帧率30 fps,无音频
- observation.images.front(视频, 480x640x3):
- 编码格式av1,像素格式yuv420p,帧率30 fps,无音频
- action(
float32, 形状 [6]):- shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
- timestamp(
float32, 形状 [1]) - frame_index(
int64, 形状 [1]) - episode_index(
int64, 形状 [1]) - index(
int64, 形状 [1]) - task_index(
int64, 形状 [1])
其他说明
- 可在此可视化数据集:https://huggingface.co/spaces/lerobot/visualize_dataset?path=anhquoc29/stack_3_box_30x3030
- 引用信息未提供(见BibTeX)



