Use 0.6 - 0.8 for best results, 1 works as well.

All donations and buzz will be used to create more models. Don't hesitate to use the comment section to recommend LORAs to train.