Diffusion model
A generative model that creates images, audio or video by learning to remove noise step by step.
Diffusion models start from random noise and iteratively denoise it toward an output that matches a prompt. Many AI image and music generators use them.
Related: Multimodal