xarm_lift_medium
收藏Hugging Face2025-02-21 更新2025-04-08 收录
下载链接:
https://huggingface.co/datasets/lerobot/xarm_lift_medium
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是使用LeRobot创建的机器人数据集,包含800个episodes、20000帧、1个任务和800个视频。数据集结构包括观察(图像和状态)、动作以及其他元数据,如episode和帧索引、时间戳、奖励和完成标志。数据集与题为《Temporal Difference Learning for Model Predictive Control》的论文相关,并采用MIT许可证。
This is a robotic dataset created using LeRobot, which contains 800 episodes, 20000 frames, 1 task, and 800 videos. The dataset structure includes observations (images and states), actions, and additional metadata such as episode and frame indices, timestamps, rewards, and termination flags. This dataset is associated with the paper titled "Temporal Difference Learning for Model Predictive Control" and is licensed under the MIT License.
提供机构:
lerobot创建时间:
2024-03-25
原始信息汇总
数据集概述
基本信息
- 名称: xarm_lift_medium
- 主页: https://www.nicklashansen.com/td-mpc/
- 论文: https://arxiv.org/abs/2203.04955
- 许可证: MIT
- 任务类别: 机器人学
- 标签: LeRobot
数据集结构
- 总集数: 800
- 总帧数: 20000
- 总任务数: 1
- 总视频数: 800
- 总块数: 1
- 块大小: 1000
- 帧率: 15 FPS
- 分割: 训练集 (0:800)
数据路径
- 数据文件:
data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet - 视频文件:
videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4
特征
- observation.image:
- 类型: 视频
- 形状: [84, 84, 3]
- 视频信息:
- 帧率: 15.0
- 编解码器: av1
- 像素格式: yuv420p
- 深度图: 否
- 音频: 无
- observation.state:
- 类型: float32
- 形状: [4]
- 名称: motor_0, motor_1, motor_2, motor_3
- action:
- 类型: float32
- 形状: [4]
- 名称: motor_0, motor_1, motor_2, motor_3
- 其他特征:
- episode_index (int64)
- frame_index (int64)
- timestamp (float32)
- next.reward (float32)
- next.done (bool)
- index (int64)
- task_index (int64)
引用
bibtex @inproceedings{Hansen2022tdmpc, title={Temporal Difference Learning for Model Predictive Control}, author={Nicklas Hansen and Xiaolong Wang and Hao Su}, booktitle={ICML}, year={2022} }



