This version is trained with only 15 images instead of 50, but manages to achieve a result more in acurrate with the artist's most recent style <3
Model created by myself <3
