lzzcd001/nabla-gfn

Official Implementation of Nabla-GFlowNet (ICLR 2025)

21
/ 100
Experimental

This tool helps researchers and AI practitioners fine-tune existing text-to-image diffusion models like Stable Diffusion. It takes a pre-trained model and a reward signal (e.g., image aesthetics or human preference scores) and outputs a refined model that generates higher-quality or more relevant images while maintaining the original model's diversity. It's for anyone building or deploying AI models who needs to adapt a general image generation model to specific quality standards or creative briefs.

No commits in the last 6 months.

Use this if you need to improve the output quality or align an existing diffusion model's image generation with specific preferences or a reward function, without losing the model's ability to produce diverse results.

Not ideal if you are looking for a tool to train a diffusion model from scratch or if you need to fine-tune a model for tasks completely unrelated to image generation.

AI-model-finetuning image-generation generative-AI AI-research diffusion-models
No License Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 7 / 25
Maturity 8 / 25
Community 4 / 25

How are scores calculated?

Stars

28

Forks

1

Language

Python

License

Last pushed

May 03, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/lzzcd001/nabla-gfn"

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