Architecture
Diffusion Model
Generative model that learns to reverse a gradual noising process to synthesise data.
Definition
Diffusion models add Gaussian noise to training data and learn to reverse it step-by-step. Sampling starts from pure noise and iteratively denoises into images, audio or video. Stable Diffusion, FLUX, Imagen, Veo and Sora are all diffusion-based generative systems.
Common use cases
- Image generation
- Video generation
- Audio synthesis
Related models
stable-diffusion
flux
imagen
sora