遇见数据集

so101_pick_snickers_front1280x720_v1_20260624_204224

收藏
Hugging Face2026-06-24 更新2026-06-24 收录
官方服务:

资源简介:

该数据集是一个机器人操作数据集,由LeRobot创建,用于机器人学任务。数据集包含50个总episodes和22844个总帧,数据以parquet格式存储,视频文件以mp4格式存储,分辨率为1280x720,帧率为30fps。数据集特征包括动作(包含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(同样包含6个关节位置)、前端图像观察(RGB视频)、时间戳、帧索引、episode索引、索引和任务索引。机器人类型为so_follower,可能用于模仿学习或强化学习应用,如物体抓取或操作任务。数据集总数据文件大小为100MB,视频文件大小为200MB,分割为训练集(包含所有50个episodes)。

This is a robotic manipulation dataset created by LeRobot for robotics research tasks. It contains a total of 50 episodes and 22844 frames. The dataset data is stored in Parquet format, while the accompanying video files are in MP4 format, with a resolution of 1280×720 and a frame rate of 30 fps. The dataset includes the following features: actions (comprising 6 joint positions: shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation, and gripper position), observation states (also containing 6 joint positions), front-facing RGB video image observations, timestamps, frame indices, episode indices, sample indices, and task indices. The target robot type is 'so_follower', and the dataset can be utilized for applications such as imitation learning or reinforcement learning, including object grasping and manipulation tasks. The total size of the dataset's data files is 100 MB, and the video files have a total size of 200 MB. The dataset is split into a training set that includes all 50 episodes.

提供机构:
Steveee7
创建时间:
2026-06-24
原始信息汇总

数据集概述

数据集名称:so101_pick_snickers_front1280x720_v1_20260624_204224

许可证:Apache-2.0

任务类别:机器人学 (Robotics)

标签:LeRobot

创建工具:使用 LeRobot 创建


数据集结构

  • 编码版本:v3.0
  • 帧率:30 FPS
  • 总片段数:50
  • 总帧数:22,844
  • 总任务数:1
  • 数据块大小:1000
  • 数据文件大小:100 MB
  • 视频文件大小:200 MB
  • 机器人类型:so_follower
  • 数据分割:训练集(片段 0:50)

数据路径

  • 数据文件:data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
  • 视频文件:videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4

特征字段

字段 数据类型 形状 描述
action float32 [6] 动作:包含6个关节位置(shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper)
observation.state float32 [6] 观测状态:与动作同义的6个关节位置
observation.images.front video [720, 1280, 3] 前置摄像头视频:分辨率1280x720,RGB三通道,AV1编码,30 FPS
timestamp float32 [1] 时间戳
frame_index int64 [1] 帧索引
episode_index int64 [1] 片段索引
index int64 [1] 全局索引
task_index int64 [1] 任务索引(当前仅1个任务)

引用信息

目前无可用引用信息(BibTeX 待补充)。

二维码
社区交流群
二维码
科研交流群
商业服务