SO101_ACT-recordtest

该数据集是使用LeRobot创建的，主要用于机器人技术领域。数据集包含24个片段，总计16,746帧，存储在parquet文件中，并配有相关的视频文件。数据集的结构包括动作、观察状态、来自手腕和顶部视图的图像、时间戳和各种索引。数据集的特征包括6个浮点型动作（如肩部平移、肩部提升、肘部弯曲等）、观察状态（与动作相同的6个浮点型数据）、来自手腕和顶部视图的视频图像（分辨率为480x640，3通道，30帧/秒），以及其他索引信息。数据集的许可证为Apache-2.0。

This dataset was constructed using LeRobot and is primarily designed for robotics research. It contains 24 segments totaling 16,746 frames, stored in Parquet files paired with associated video files. The dataset's structure includes actions, observation states, images captured from the wrist-mounted camera and top-down perspective, timestamps, and multiple indices. Its featured components consist of 6-dimensional floating-point actions (e.g., shoulder translation, shoulder elevation, elbow flexion, and others), observation states (6-dimensional floating-point data consistent with the actions), video frames from the wrist and top-down views with a resolution of 480×640, 3 color channels, and 30 frames per second, alongside other index-related information. The dataset is licensed under Apache-2.0.