aperture-v2-style

This model try to replicate the style of the lexica aperture v2 model.

It was train using about 5000 portrait images and prompts scrapped from the lexica API

V1.1 is trained after manually curating the same image set to keep only the best 3000 images