Text-to-Image Diffusion Models Cannot Count, and Prompt Refinement Cannot Help

Open in new window