Loading…
Glossary term
A generative model that creates images or video by gradually removing noise.
A diffusion model learns to reverse a noising process. During training, real images are gradually corrupted with noise and the network learns to undo it. To create something new it starts from random noise and removes it step by step, guided by a text prompt. Diffusion now powers most image and video generation.
It replaced GANs as the leading approach for realistic generated images, video and audio.