#71 · Primary category: AI Design & Prototyping

pix2code

datasets deep-learning deep-neural-networks front-end-development graphical-user-interface

pix2code: Generating Code from a Graphical User Interface Screenshot

Project last updated:05/29/24

GitHub Stars

12.0K

Forks

1.4K

Contributors

5

License

Apache-2.0

Why we included this project

pix2code dates from 2017, when Tony Beltramelli showed that a neural network could learn to turn a GUI screenshot into front-end code. The model is trained end to end on screenshots and outputs a domain-specific language that gets compiled into Android XML, iOS storyboards, or Bootstrap-style HTML and CSS, reaching over 77% accuracy across the three platforms. For anyone studying how vision and code generation intersect, it is a useful reference precisely because it is so self-contained: the source and datasets ship with the repo, so you can re-run the original experiments or build on them. Just don't mistake it for a production tool. The author says plainly that this is an educational research project, not something to wire into a real build pipeline.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category