fold_towel_20260707_10_20260707_182627

Name: fold_towel_20260707_10_20260707_182627
Creator: access-robotics-team
Published: 2026-07-07T17:36:30+08:00

Hugging Face2026-07-07 更新2026-07-07 收录

机器人折叠毛巾

数据链接：

https://huggingface.co/datasets/access-robotics-team/fold_towel_20260707_10_20260707_182627 数据链接链接失效反馈

官方服务：

资源简介：

该数据集是一个用于机器人折叠毛巾任务的数据集，使用LeRobot工具创建。数据集包含双手机器人（bi_openarm_follower）的动作和状态观测数据，动作数据包括左右机械臂的16个关节位置（含夹爪），状态观测同样包含16个关节位置。此外，数据集提供了多视角图像观测：顶部摄像头、右腕摄像头和左腕摄像头，每帧图像分辨率为480x640，3通道，视频编码为av1，帧率为30fps。数据集还包括时间戳、帧索引、episode索引、任务索引等元数据。总共有10个episodes，11549帧，数据以parquet文件和视频文件形式存储，总数据量约300MB（数据文件100MB，视频文件200MB）。数据集适用于机器人学习、模仿学习或强化学习任务。

This dataset is developed for robotic towel folding tasks, and was created using the LeRobot toolkit. It contains action and state observation data from the dual-arm robot (bi_openarm_follower): the action data includes 16 joint positions (including grippers) for both the left and right robotic arms, while the state observation data also covers 16 joint positions. Additionally, the dataset provides multi-view image observations captured by a top-mounted camera, a right-wrist camera, and a left-wrist camera. Each image frame has a resolution of 480x640 with 3 color channels, uses AV1 video encoding, and runs at 30 fps. The dataset also includes metadata such as timestamps, frame indices, episode indices, and task indices. In total, there are 10 episodes and 11549 frames. The data is stored in Parquet files and video files, with an overall size of approximately 300 MB (100 MB for data files and 200 MB for video files). This dataset is suitable for robotic learning, imitation learning, and reinforcement learning tasks.

提供机构：

access-robotics-team

创建时间：

2026-07-07

原始信息汇总

数据集概述：fold_towel_20260707_10_20260707_182627

许可证：Apache-2.0
任务类别：机器人学（Robotics）
标签：LeRobot

数据集结构

数据格式：Parquet 文件，存放于 data/*/*.parquet 路径下
总帧数：11,549 帧
总片段数：10 个片段
总任务数：1 个任务
帧率：30 FPS
数据集文件大小：100 MB
视频文件大小：200 MB
机器人类型：bi_openarm_follower

数据特征

特征名称	数据类型	形状	说明
`action`	float32	(16,)	包含左右臂各7个关节位置及1个夹爪位置
`observation.state`	float32	(16,)	与`action`相同的机器人状态信息
`observation.images.top`	video	(480, 640, 3)	顶部摄像头视频，AV1 编码，30 FPS
`observation.images.right_wrist`	video	(480, 640, 3)	右手腕摄像头视频，AV1 编码，30 FPS
`observation.images.left_wrist`	video	(480, 640, 3)	左手腕摄像头视频，AV1 编码，30 FPS
`timestamp`	float32	(1,)	时间戳
`frame_index`	int64	(1,)	帧序号
`episode_index`	int64	(1,)	片段序号
`index`	int64	(1,)	索引
`task_index`	int64	(1,)	任务序号