Towards Training One-Step Diffusion Models Without Distillation

Open in new window