utokyo_saytap

该数据集使用LeRobot创建，包含机器人相关的视频和传感器数据。数据集结构详细描述了数据文件、视频文件以及各种特征，如观察、动作和时间戳。数据集与机器人学相关，包括视频和传感器数据。引用信息链接到一篇题为'SayTap: Language to Quadrupedal Locomotion'的论文。

This dataset was created using LeRobot, and contains robot-related video and sensor data. The dataset structure elaborately describes data files, video files, and various features such as observations, actions, and timestamps. This dataset is relevant to robotics, and includes video and sensor data. The citation information links to a paper titled 'SayTap: Language to Quadrupedal Locomotion'.

原始信息汇总

数据集概述

基本信息

• 名称: lerobot/utokyo_saytap
• 主页: https://saytap.github.io/
• 论文: https://arxiv.org/abs/2306.07580
• 许可证: MIT
• 任务类别: 机器人学
• 标签: LeRobot

数据集结构

• 总集数: 20
• 总帧数: 22937
• 总任务数: 20
• 总视频数: 40
• 总块数: 1
• 块大小: 1000
• 帧率: 5 FPS
• 数据路径: data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet
• 视频路径: videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4

特征

1. 观察图像:
   • observation.images.wrist_image 和 observation.images.image
     • 数据类型: 视频
     • 形状: [64, 64, 3]
     • 视频信息:
       • 帧率: 5.0
       • 编解码器: av1
       • 像素格式: yuv420p
       • 深度图: 否
       • 音频: 无
2. 语言指令:
   • language_instruction
     • 数据类型: 字符串
     • 形状: [1]
3. 观察状态:
   • observation.state
     • 数据类型: float32
     • 形状: [30]
     • 名称: 30个电机状态
4. 动作:
   • action
     • 数据类型: float32
     • 形状: [12]
     • 名称: 12个电机动作
5. 其他特征:
   • timestamp, episode_index, frame_index, next.reward, next.done, index, task_index

引用

bibtex @article{saytap2023, author = {Yujin Tang and Wenhao Yu and Jie Tan and Heiga Zen and Aleksandra Faust and Tatsuya Harada}, title = {SayTap: Language to Quadrupedal Locomotion}, eprint = {arXiv:2306.07580}, url = {https://saytap.github.io}, note = {https://saytap.github.io}, year = {2023} }