five

record-test_20260708_184527

收藏
Hugging Face2026-07-08 更新2026-07-08 收录
下载链接:
https://huggingface.co/datasets/Yamaguchi-TR/record-test_20260708_184527
下载链接
链接失效反馈
官方服务:
资源简介:
这是一个用于机器人任务的数据集,基于LeRobot创建。数据集包含5个训练集片段,总帧数为2498帧,数据来源于so_follower机器人。其关键特征包括:动作(由6个关节位置组成,涉及肩部平移、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(与动作相同的6个关节位置)、两个摄像头视频观测(顶部摄像头和夹爪摄像头,分辨率为360x640,帧率为30fps,采用H.264编码和RGB格式),以及元数据如时间戳、帧索引、片段索引和任务索引。数据以Parquet文件格式存储,视频文件单独存放,总数据大小约为100MB,视频大小约为200MB。该数据集适用于机器人控制和学习任务。

This is a robotic task-oriented dataset developed based on LeRobot. The dataset contains 5 training episodes with a total of 2498 frames, and the data is collected from the so_follower robot. Its key features include: actions (composed of 6 joint positions, covering shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation and gripper position), state observations (the same 6 joint positions as the actions), two camera video observations (top camera and gripper camera, with a resolution of 360x640, a frame rate of 30fps, using H.264 encoding and RGB format), as well as metadata such as timestamps, frame indices, episode indices and task indices. The data is stored in Parquet file format, while the video files are stored separately. The total size of the dataset is approximately 100 MB, and the total size of the video files is about 200 MB. This dataset is suitable for robotic control and learning tasks.
提供机构:
Yamaguchi-TR
创建时间:
2026-07-08
原始信息汇总

数据集概述

  • 名称: record-test_20260708_184527
  • 来源: 使用 LeRobot 框架创建
  • 许可证: Apache-2.0
  • 任务类别: 机器人学 (robotics)
  • 标签: LeRobot

数据集结构

  • 总片段数 (episodes): 5
  • 总帧数 (frames): 2498
  • 总任务数: 1
  • 帧率 (FPS): 30
  • 机器人类型: so_follower

特征 (Features)

特征名称 数据类型 形状 描述
action float32 (6,) 包含6个关节动作:shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
observation.state float32 (6,) 机器人状态,与动作名称相同的6个关节位置
observation.images.gripper video (H264) (360, 640, 3) 夹爪视角的RGB视频,高度360,宽度640,3通道,帧率30fps
observation.images.top video (H264) (360, 640, 3) 顶部视角的RGB视频,高度360,宽度640,3通道,帧率30fps
timestamp float32 (1,) 时间戳
frame_index int64 (1,) 帧索引
episode_index int64 (1,) 片段索引
index int64 (1,) 索引
task_index int64 (1,) 任务索引

数据划分

  • 训练集 (train): 包含所有5个片段 (episodes 0 到 4)

文件存储

  • 数据文件: data/*/*.parquet (Parquet格式,约100 MB)
  • 视频文件: videos/*/chunk-*/file-*.mp4 (MP4格式,H264编码,约200 MB)
  • 数据分块大小 (chunks_size): 1000 帧/块
二维码
社区交流群
二维码
科研交流群
商业服务