this is my first attempt at making a lora that I saw this place didn't have. it was trained on 20 pictures I was able to find. it generally captures the face pretty well but I have mixed results on the rest of the body, but it seemed decent enough.