Asked about Generating by removing noise
Why did diffusion displace GANs?
An answer that works
Training stability. A GAN's adversarial objective suffers mode collapse and oscillation, while diffusion has a simple regression objective that trains reliably and scales predictably. The cost is much slower sampling.
Reading an answer is not being able to give one. The section that teaches this is Generating by removing noise, 4 hours of Diffusion & generative models.
Learn this properlyAlso asked about Diffusion & generative models
How do diffusion models work?What is latent diffusion?How do you trade quality against latency in a diffusion model?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?