DPO (Direct Preference Optimization) LoRA for XL and 1.5 - OpenRail++ - SDXL - V1.0 Showcase