用戶LoRA
smokinglora
This model was trained on 40 images of people, smoking cigarettes. Its far from perfect, but still a major improvement to prompt only. works best around 0,8-1
