Dataset · OpenGVLab
Dataset for training agents in graphical user interfaces
OpenGVLab/GUI-Odyssey
English dataset with screenshots, text, and tabular data from OpenGVLab for GUI navigation.
What's inside
The dataset includes screen images, text, and tabular data in English, collected by the OpenGVLab team. This version contains between 1 and 10 thousand records, and the authors warn in the card about a newer version of the repository. The dataset is accompanied by a research paper and open-source code on GitHub.
What you can do with it
- Training multimodal UI agents for automating on-screen actions
- Training VLMs to recognize interface elements and execute user scenarios
- Evaluating app navigation models (it is better to use the updated version from the link in the card)
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 67K
- Likes
- 26
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("OpenGVLab/GUI-Odyssey")