Hacker News new | ask | show | jobs
by kadushka 469 days ago
No, we typically train any diffusion model on a single step (randomly chosen).