Lora trained on a dataset of 30+ high-resolution (1024x1024) images of Michael Jordan the goat, specifically for use for his face. It will not give a style or pose!