The LORA model, a woman modeled as a photo model, influencer, or simply a beautiful, shapely woman, was trained on 45 generated high-resolution images. I tried to select images as close as possible to the adopted visual concept, so that the model maintained a relatively uniform and repeatable image. The vast majority of the images used for training were non-nudity, with the exception of a few images with partial nudity, which were added to teach the model the exact female figure I intended as the target.