Self-Play Fine-Tuning of Diffusion Models for Text-to-Image Generation Paper โข 2402.10210 โข Published Feb 15 โข 28