eren23/one_layer_image_gen

A PyTorch implementation replicating the core logic of FAE (Feature Auto-Encoder) from the paper "One Layer Is Enough: Adapting Pretrained Visual Encoders for Image Generation" (arXiv:2512.07829).

16
/ 100
Experimental

This project helps machine learning researchers adapt powerful existing image understanding models to generate new images. It takes the detailed numerical representations (features) from a model like DINOv3 and transforms them into a compact, generative format, which then produces realistic new images. This tool is for researchers focused on computer vision and generative AI.

Use this if you are a machine learning researcher aiming to build image generation models more efficiently by leveraging and adapting state-of-the-art visual feature encoders, rather than training from scratch.

Not ideal if you are looking for an out-of-the-box image generation tool for end-user applications or if you require extremely high-resolution image outputs beyond 224x224.

generative-ai computer-vision deep-learning-research image-synthesis
No License No Package No Dependents
Maintenance 6 / 25
Adoption 5 / 25
Maturity 5 / 25
Community 0 / 25

How are scores calculated?

Stars

12

Forks

Language

Jupyter Notebook

License

Last pushed

Dec 20, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/generative-ai/eren23/one_layer_image_gen"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.