Another of my earliest attempts at training a LoRA. It has issues that I do not yet know how to fix. The biggest issue is the model training data and available material, I would assume, but the time/skill investment to get a better result isn't something I have currently.

I typically slap this together with some different Style and Concept LoRA's at 896x1152 to achieve cultural enrichment. Not sure what the sweet spot is on other settings.

Hope this is to the joy of some, and if I keep with training models I will surely revisit this to get better results.

Based on CG from the game.