addtt/variational-diffusion-models

PyTorch implementation of Variational Diffusion Models.

37
/ 100
Emerging

This project offers a clear implementation of Variational Diffusion Models, primarily focused on accurately estimating the likelihood of images rather than just their visual quality. It takes raw image data (like the CIFAR10 dataset) and provides a model that can generate new images, along with metrics like the variational lower bound (VLB) to quantify how well the model understands the underlying data distribution. This is intended for researchers and machine learning practitioners who need to rigorously evaluate generative models based on their probabilistic fidelity.

104 stars. No commits in the last 6 months.

Use this if you are a researcher or practitioner interested in understanding and implementing Variational Diffusion Models with an emphasis on likelihood estimation and educational clarity, especially for image generation tasks.

Not ideal if your primary goal is to generate visually stunning, high-fidelity images without a deep concern for the underlying probabilistic modeling or if you need advanced features like variance minimization with the γη network.

generative-modeling image-synthesis probabilistic-modeling machine-learning-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 12 / 25

How are scores calculated?

Stars

104

Forks

10

Language

Python

License

MIT

Last pushed

Apr 23, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/addtt/variational-diffusion-models"

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