Asked about Train a tiny diffusion model
You trained a diffusion model and samples look like blurry blobs. What might be wrong?
An answer that works
Too little training, a network too small for the data, a wrong noise schedule so the final step is not near pure noise, a bug in the sampling formula, or the time step not reaching the network properly. Checking that the forward process at the last step is indistinguishable from noise and plotting intermediate samples narrow it down fast.
Reading an answer is not being able to give one. The section that teaches this is Train a tiny diffusion model, 5 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?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 need image generation under one second per image. What are your options?