Term · Advanced
Diffusion Model
The AI architecture behind most image and video generators, it denoises randomness into a picture.
A diffusion model generates images by starting from pure noise and gradually removing it, step by step, until a coherent picture matching your prompt emerges. It's the engine under Stable Diffusion, Midjourney, and most modern image and video tools, which is why understanding it helps you predict how settings like steps and guidance behave. Knowing the mechanism makes prompt and parameter tuning far less like guesswork.
Also called Diffusion, Latent Diffusion
Related
- Generative AIAI that produces new content, images, video, audio, text, from patterns it learned.
- Text-to-ImageGenerating still images from a written prompt using AI.
- Checkpoint ModelA trained base model file that defines an AI image generator's overall style and knowledge.
- SeedThe starting number that determines an AI image's random noise, reuse it to reproduce results.
- Negative PromptA list of things you tell an AI image model to avoid or exclude from the result.