26/03/01 V1Despite the name the lora is generally good at any art. The noticeable thing is that almost all faces generated with this model are anime-like, with certain style of eyes and hair.Training is done with a lot of 768x768 images which is a great success since before I've had only 1024x1024 trainings or unsuccessful mixed-resolution one. This is the proof of concept for me. You can generate whatever resolution you'd like.

Files with SHA256: 36316e6bb0ec0d152d5719f0f9944c0f555edc8eab6322c50ed06621aa092296