five

rollout_so101_pick_test4_20260608_005456

收藏
Hugging Face2026-06-08 更新2026-06-08 收录
下载链接:
https://huggingface.co/datasets/2006gk/rollout_so101_pick_test4_20260608_005456
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学数据集,使用LeRobot创建,专注于机器人动作和观察数据的收集。数据集包含机器人的动作(action)和观察状态(observation.state),均为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。观察数据包括正面和侧面的视频图像,分辨率为480x640,3通道(RGB),帧率为30fps,使用AV1编解码器。数据集还包含时间戳、帧索引、episode索引、任务索引等元数据,用于组织数据。总共有7个episodes,12566帧,1个任务,数据文件大小为100MB,视频文件大小为200MB,机器人类型为so_follower。数据以parquet格式存储,视频以MP4格式存储,适用于机器人控制、强化学习或计算机视觉任务。

This robotics dataset was developed using LeRobot, focusing on the collection of robot action and observation data. It includes robot actions (action) and observation states (observation.state), both 6-dimensional floating-point arrays representing robot joint positions, including shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation, and gripper position. The observation data comprises front and side-view RGB video images with a resolution of 480×640, 3 channels, a frame rate of 30 fps, and encoded with the AV1 codec. Additional metadata such as timestamps, frame indices, episode indices, and task indices are included to facilitate data organization. The dataset consists of 7 episodes, 12566 frames, and 1 task, with a structured data file size of 100 MB, a video file size of 200 MB, and utilizes the so_follower robot platform. The structured data is stored in Parquet format, while the video files are stored in MP4 format, and this dataset is applicable to robotics control, reinforcement learning, and computer vision tasks.
提供机构:
2006gk
创建时间:
2026-06-08
原始信息汇总

数据集概述

  • 数据集名称: rollout_so101_pick_test4_20260608_005456
  • 创建者: 2006gk
  • 许可证: Apache-2.0
  • 任务类别: 机器人 (Robotics)
  • 标签: LeRobot
  • 创建工具: 使用 LeRobot 创建

数据集结构

  • 代码库版本: v3.0
  • 帧率 (FPS): 30
  • 总片段数: 7
  • 总帧数: 12,566
  • 总任务数: 1
  • 机器人类型: so_follower
  • 数据文件格式: Parquet
  • 视频文件格式: MP4 (编码: av1)
  • 数据文件大小: 约 100 MB
  • 视频文件大小: 约 200 MB
  • 数据路径模式: data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
  • 视频路径模式: videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
  • 数据划分: 全部数据用于训练 (train: "0:7")

特征 (Features)

特征名 数据类型 形状 描述
action float32 [6] 6维动作:shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
observation.state float32 [6] 6维观察状态(同动作维度)
observation.images.front 视频 (AV1) [480, 640, 3] 前视摄像头图像 (480x640, 30fps, 3通道)
observation.images.side 视频 (AV1) [480, 640, 3] 侧视摄像头图像 (480x640, 30fps, 3通道)
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

配置

  • 配置名称: default
  • 数据文件: data/*/*.parquet
二维码
社区交流群
二维码
科研交流群
商业服务