# ScreenSpot-v2-variaints

<div style="display: flex; gap: 10px;">
  <a href="https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding">
    <img src="https://img.shields.io/badge/github-%23121011.svg?style=for-the-badge&logo=github&logoColor=white" alt="github" class="img-fluid" />
  </a>
  <a href="https://likaixin2000.github.io/papers/ScreenSpot_Pro.pdf">
    <img src="https://img.shields.io/badge/arXiv-paper-b31b1b.svg?style=for-the-badge" alt="paper" class="img-fluid" />
  </a>
  <a href="https://gui-agent.github.io/grounding-leaderboard/screenspot">
    <img src="https://img.shields.io/badge/Leaderboard-8A2BE2?style=for-the-badge" alt="paper" class="img-fluid" />
  </a>
</div>

The [ScreenSpot dataset](https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2) with 4 types of instructions: 
  - `instruction`: original instruction from ScreenSpot,
  - `action`: clarifies the action to take,
  - `description`: describes the target UI element.
  - `negative`: an operation that can not be done in the screenshot.

Compatible code can be found in the GitHub repo above.