Asked about Using diffusion models in practice
How do you trade quality against latency in a diffusion model?
An answer that works
Step count, primarily, since latency scales roughly with it. Modern samplers reach good quality in around twenty steps, and distilled models go much lower at a real quality cost. Guidance strength is the second dial.
Reading an answer is not being able to give one. The section that teaches this is Using diffusion models in practice, 4 hours of Diffusion & generative models.
Learn this properlyAlso asked about Diffusion & generative models
How do diffusion models work?Why did diffusion displace GANs?What is latent diffusion?What is classifier-free guidance?How would you add a specific style to a diffusion model?Write down the diffusion training objective and explain it.You trained a diffusion model and samples look like blurry blobs. What might be wrong?You need image generation under one second per image. What are your options?