Denoising Diffusion Models

Denoising Diffusion Model

Generative architecture where data is progressively corrupted by adding Gaussian noise, then restored by learning to reverse this diffusion process. These models excel at generating high-quality images by iteratively denoising random samples.

← Back