Did this initially just to see if training a model on a more simplistic style would work well, but I ended up pretty happy with the results enough to publish it. Planning on possibly adding more to it's dataset/training it further Changes: - Added several of their OC's. Blok, Nox, Couch Goblin

Smokey Blokey Style

I went back and re-tagged every training image from the first version, and added a bunch more. I decent amount of inflation/expansion stuff was also added to the data to see how it handles that with the style-

Files with SHA256: b8b2a4493c47f8bcb11541056d166dd29a674de7831f7a703e122cee4f74ad80