swyoon/Diffusion-by-MaxEntIRL

The official repository for NeurIPS 2024 Oral

36
/ 100
Emerging

This project offers a refined approach for researchers and engineers working with advanced image generation models. It takes existing diffusion models (like DDPM, DDGAN, or EDM) and, using a maximum entropy inverse reinforcement learning technique, enhances their image generation capabilities. The output is a new, more performant generative model capable of producing high-quality images from datasets like CIFAR-10, ImageNet 64x64, and LSUN Bedroom.

No commits in the last 6 months.

Use this if you are a machine learning researcher or practitioner looking to improve the fidelity and performance of diffusion-based image generation models for common image datasets.

Not ideal if you are looking for an out-of-the-box solution for general content creation or if you lack expertise in deep learning model training and evaluation.

deep-learning-research image-generation generative-models computer-vision machine-learning-engineering
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 13 / 25

How are scores calculated?

Stars

32

Forks

5

Language

Python

License

MIT

Last pushed

Mar 20, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/swyoon/Diffusion-by-MaxEntIRL"

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