LoRA in Granblue Fantasy style. These are still the first test versions while I’m searching for the most suitable settings. I hope for your understanding.

V0.7 T2

This is not a new version, but rather a different training setup on the same dataset, which produces a different result. No trigger word is required for this version.

V0.7T

The LoRA is based on Janku v5. Works with Noob and Illustrious.