DDPM

Denoising diffusion probabilistic models (DDPM) are a class of generative models that produce data by reversing a gradual noising process. They have gained prominence for their ability to generate high-quality images and other complex data.