Diffusion Model Alignment Using Direct Preference Optimization | Read Paper on Bytez