leokswang/23122025-foldclo-01_lerobot_format
收藏资源简介:
该数据集是一个机器人学数据集,使用LeRobot工具创建,专为YAM机器人设计。它包含9个episodes,总计16772帧,帧率为30fps,数据以parquet格式存储。数据集提供了多模态观测数据,包括来自三个摄像头(顶部、左侧、右侧)的图像(每个图像尺寸为360x640x3),以及14维的机器人状态观测。此外,还包含14维的动作数据、时间戳、帧索引、episode索引等元数据。该数据集适用于机器人控制任务的训练和评估,支持强化学习或模仿学习应用。
This robotics dataset was developed using the LeRobot toolkit and is specifically tailored for the YAM robot. It consists of 9 episodes with a total of 16772 frames at a frame rate of 30 fps, and the data is stored in Parquet format. The dataset provides multimodal observation data, including images from three cameras (top, left, right) with a resolution of 360×640×3 per image, as well as 14-dimensional robot state observations. Additionally, it also contains 14-dimensional action data, timestamps, frame indices, episode indices and other metadata. This dataset is suitable for training and evaluation of robotic control tasks, and supports applications of reinforcement learning or imitation learning.



