m2m_smolvla_finetune_dataset

该数据集是通过LeRobot创建的机器人相关数据集，包含20个episodes，15132帧数据，1个任务。数据集结构包括动作（6个浮点数表示机器人关节位置）、观察状态（6个浮点数表示机器人关节位置）、前端图像（1080x1920x3的视频帧）、时间戳、帧索引、episode索引等特征。数据以parquet格式存储，视频以mp4格式存储。

This robotics-related dataset was created via LeRobot. It contains 20 episodes, 15,132 frames, and 1 single task. The dataset structure includes features such as actions (6 floating-point numbers representing robot joint positions), observation states (6 floating-point numbers representing robot joint positions), front-facing images (1080×1920×3 video frames), timestamps, frame indices, episode indices, and other relevant attributes. The structured data is stored in Parquet format, while the videos are stored in MP4 format.