eval_taken_11

该数据集由LeRobot创建，专注于机器人学领域，特别是移动AI机器人的操作任务。数据集包含一个完整的episode，总共有2310帧，对应一个特定任务。数据以parquet文件格式存储，视频以MP4格式存储，帧率为30fps。数据集的特征包括机器人的动作（如16个关节位置和速度）、观测状态（同样包含16个关节位置和速度），以及来自三个摄像头的图像观测：一个高角度摄像头（cam_high）和两个手腕摄像头（cam_left_wrist和cam_right_wrist），每个摄像头的视频分辨率为480x640，颜色通道为3。此外，数据集还包括时间戳、帧索引、episode索引、索引和任务索引等元数据。这些数据可用于机器人控制、视觉导航和机器学习模型的训练与评估。This dataset was developed by LeRobot, focusing on the field of robotics, particularly manipulation tasks for mobile AI robots. It contains one complete episode with a total of 2310 frames corresponding to a specific task. The structured data is stored in Parquet file format, while the associated videos are saved in MP4 format at 30 frames per second (fps). The dataset features include robot actions (such as 16 joint positions and velocities), observation states (also containing 16 joint positions and velocities), and image observations from three cameras: one high-angle camera (cam_high) and two wrist cameras (cam_left_wrist and cam_right_wrist). Each camera’s video has a resolution of 480×640 with 3 color channels. Additionally, the dataset includes metadata such as timestamps, frame indices, episode indices, sample indices, and task indices. These data can be used for robot control, visual navigation, and the training and evaluation of machine learning models.