This is a LORA trained to augment the capabilities of SDXL. SDXL already produces very good quality images, this LORA is intended to exaggerate the current style. Recommended Lora weights are - (Model-1) (Clip-0.25)