benediktfesl/Diffusion_MSE

Implementation of the paper "On the Asymptotic Mean Square Error Optimality of Diffusion Models."

29
/ 100
Experimental

This project provides an implementation to reproduce research results on the asymptotic mean square error optimality of diffusion models, especially for denoising tasks. It takes various data types like image data (e.g., MNIST, Fashion MNIST) or audio data, applies denoising using diffusion models or Gaussian Mixture Model (GMM) baselines, and outputs evaluation metrics. This tool is designed for researchers in machine learning or signal processing who are investigating the theoretical properties and performance of diffusion models for denoising.

No commits in the last 6 months.

Use this if you are a researcher who needs to reproduce or validate the simulation results related to the asymptotic mean square error optimality of diffusion models for denoising tasks.

Not ideal if you are looking for a user-friendly application to simply apply denoising to your own custom images or audio without a research focus on model optimality.

machine-learning-research signal-denoising diffusion-models statistical-modeling computational-statistics
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 5 / 25
Maturity 16 / 25
Community 6 / 25

How are scores calculated?

Stars

14

Forks

1

Language

Python

License

BSD-3-Clause

Last pushed

May 14, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/benediktfesl/Diffusion_MSE"

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