five

so101-fold-towel-2camera_2_20260630_204225

收藏
Hugging Face2026-06-30 更新2026-06-30 收录
下载链接:
https://huggingface.co/datasets/aHiroakiIshikawa/so101-fold-towel-2camera_2_20260630_204225
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工具创建,专注于机器人操作任务。数据集包含10个episodes,总共6309帧,涉及1个任务。数据特征包括:机器人动作(6个关节的位置,如肩部平移、肩部升降、肘部弯曲、手腕弯曲、手腕旋转和夹爪位置),观察状态(同样为6个关节位置),以及两个摄像头的视频观察(手腕摄像头和顶部摄像头),视频分辨率为480x640,帧率为30fps,格式为AV1编码。此外,数据集还包含时间戳、帧索引、episode索引、任务索引等信息。机器人类型为so_follower。数据集以Parquet格式存储,总大小约为300MB(数据文件100MB,视频文件200MB),用于训练和机器人控制研究。

This dataset is a robotics dataset created using the LeRobot tool, focusing on robot manipulation tasks. It contains 10 episodes, totaling 6309 frames, and involves 1 task. The data features include: robot actions (positions of 6 joints, such as shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position), observation states (similarly 6 joint positions), and video observations from two cameras (wrist camera and top camera), with a resolution of 480x640, frame rate of 30fps, and AV1 encoding format. Additionally, the dataset includes metadata such as timestamps, frame indices, episode indices, and task indices. The robot type is so_follower. The dataset is stored in Parquet format, with a total size of approximately 300MB (100MB for data files and 200MB for video files), and is intended for training and robot control research.
提供机构:
aHiroakiIshikawa
创建时间:
2026-06-30
原始信息汇总

数据集概述

  • 数据集名称:so101-fold-towel-2camera_2_20260630_204225
  • 许可证:Apache-2.0
  • 任务类型:机器人学(robotics)
  • 使用框架:基于 LeRobot 构建

数据集结构

  • 代码库版本:v3.0
  • 帧率(FPS):30
  • 总片段数(Episodes):10
  • 总帧数:6309
  • 总任务数:1
  • 机器人类型:so_follower

特征(Features)

特征名称 数据类型 形状 说明
action float32 [6] 6个动作维度:shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
observation.state float32 [6] 状态观测,与动作维度对应
observation.images.wrist video(AV1编码) [480, 640, 3] 腕部摄像头图像,分辨率480x640,RGB三通道,帧率30fps
observation.images.top video(AV1编码) [480, 640, 3] 顶部摄像头图像,分辨率480x640,RGB三通道,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

数据存储

  • 数据文件:Parquet 格式,路径为 data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
  • 视频文件:MP4 格式,路径为 videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
  • 数据文件大小:约100 MB
  • 视频文件大小:约200 MB
  • 分块大小:每块1000帧

数据集划分

  • 训练集(train):包含全部10个片段(索引0至9)

可视化

可通过以下链接在线可视化数据集:
https://huggingface.co/spaces/lerobot/visualize_dataset?path=aHiroakiIshikawa/so101-fold-towel-2camera_2_20260630_204225

二维码
社区交流群
二维码
科研交流群
商业服务