first version, trained on 6 images for Illustrious
I've put some of my drawings in a small dataset to train a lora that matches my style
Use weight around 0.7
