Dataset · zonghanHZH
Web interface screenshots for training GUI agents
zonghanHZH/ShowUI-web-8k
A curated selection of 8,000 English website screenshots for training models to understand web interface elements.
What's inside
Contains 8,000 screenshots of real websites featuring diverse layouts and English UI components. This is a curated subset of a larger original dataset for the GUI grounding task (matching text to screen elements). The dataset covers a wide range of websites and is intended for pre-training multimodal models.
What you can do with it
- Pre-training vision-language models to locate buttons and controls on web pages
- LoRA fine-tuning for browser AI agents
- Evaluating recognition accuracy of web interface elements
- Tasks
- Image text → text
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 35.5K
- Likes
- 0
How to load it
from datasets import load_dataset
ds = load_dataset("zonghanHZH/ShowUI-web-8k")