This is a style model that does Moralgear style art, the artist tends to giantess, and big breasts. Tends to work best at 0.6. This model was trained on illustrious-xl and generated in Illustrious pencil-xl. Made with the Kohya tool for AI Training through google colab. Any suggestions to improve or recommended things for the data set would be appreciated.
You can find the artist here.
https://twitter.com/moralgear
Description
Now with over 300 images and retagged from top to bottom, this newest version I'm quite happy with. Since their main OC has hair flaps you may want to put 'hair flaps' in negatives to remove that type of hair. This model has much better tags, and images have been cleaned better (I've improved at that, compared to when I originally started). This version adds new images as well as more crops to bring better quality and diversity. Now optimized for Illustrious 1.0. Will probably release loras of individual OC's soon.
I'm considering taking my images from my Tworship, moralgear, ydbunny, and futon, (thats more than a 1000 images) and fusing them into a Illustrious 1.0 checkpoint. I don't know anything about checkpoint making though (and I'd still need to deep dive ydbunny and futon), but that's an idea I've been thinking about.