lucataco
/
dpo-sdxl
Direct Preference Optimization (DPO) is a method to align diffusion models to text human preferences by directly optimizing on human comparison data
Want to make some of these yourself?
Run this model