il_gym3

该数据集是通过LeRobot创建的，主要用于机器人技术领域。数据集包含5个episodes，共468帧，涉及1个任务和10个视频。数据集的结构包括观测状态（18维浮点数组）、动作（4维浮点数组）、奖励（1维浮点数组）、完成标志（布尔值）等字段。此外，数据集还包含来自前端和手腕摄像头的视频数据（128x128像素，3通道）。

This dataset was developed using LeRobot, and is primarily intended for robotics research. It consists of 5 episodes, totaling 468 frames, covering 1 task and 10 videos. The dataset structure includes fields such as observation state (18-dimensional float array), action (4-dimensional float array), reward (1-dimensional float array), and done flag (boolean value). Additionally, the dataset contains video data from the front-facing and wrist cameras, with a resolution of 128×128 pixels and 3 channels.