so100_eraser2bowl_gd2

该数据集使用LeRobot创建，主要包含机器人操作相关的数据。数据集包含40个视频片段，总计35080帧，帧率为30fps。数据集中包含6种动作特征（如主肩旋转、主肩提升等）和14种观察状态特征（如主肩旋转、主肩提升、拾取和放置坐标等）。此外，数据集还包含来自笔记本电脑的图像数据，分辨率为480x640，3通道。数据集的结构设计用于训练机器人操作任务，特别是与拾取和放置相关的任务。

This dataset was developed using LeRobot, and mainly contains data for robotic manipulation tasks. It comprises 40 video clips, totaling 35,080 frames at a frame rate of 30fps. The dataset includes 6 categories of action features (e.g., primary shoulder rotation, primary shoulder elevation, etc.) and 14 categories of observation state features (e.g., primary shoulder rotation, primary shoulder elevation, pick-and-place coordinates, etc.). Additionally, it contains image data collected from a laptop, with a resolution of 480×640 and 3 channels. The structure of this dataset is tailored for training robotic manipulation tasks, especially those related to pick-and-place operations.