Dataset · docling-project
Large UI Screenshot and Annotation Dataset
docling-project/screenparse
Over 1.4 million screenshots with detailed screen element annotations for training UI parsing models.
What's inside
The second release of a large-scale dataset for complete screen parsing, featuring 1,447,100 cleaned training screenshots across various resolutions. The data includes text and images with dense annotations of leaf interface elements to reduce noise. Created by the Docling project team for object detection and classification tasks.
What you can do with it
- Training multimodal models and GUI agents in UI element recognition
- Detecting buttons, input fields, and text blocks on screenshots for OCR and parsing
- Testing screen interaction automation systems across different resolutions
- Tasks
- Object detection, Image classification
- Size
- 1M–10M rows
- Downloads this month
- 20.4K
- Likes
- 8
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("docling-project/screenparse")