tecky-tech/Tecky-UI-Elements-VLM

Tecky UI自动化数据集是一个用于训练视觉语言模型理解桌面环境中图形用户界面(GUIs)的数据集。它包含了来自真实应用程序屏幕截图的数千个标记UI元素。数据集旨在使模型能够解析GUI屏幕截图，识别如按钮、输入框、滑块等可操作元素，根据指令预测适当的点击目标，并学习跨不同应用程序和布局的通用UI语义。

The Tecky UI Automation Dataset is a dataset for training Vision-Language Models to understand Graphical User Interfaces (GUIs) in desktop environments. It contains thousands of labeled UI elements from real-world app screenshots. The dataset is designed to enable models to parse GUI screenshots, identify actionable elements like buttons, inputs, sliders, predict appropriate click targets based on instructions, and learn generic UI semantics across different applications and layouts.