Trained using LoCon with a post-2025 dataset. This results in a slight change in style, while generalization remains excellent.
0.8-1.0trigger word:ch0u0(v0.1 can use more,even1.5 is fine)
https://twitter.com/chung0u0