sbtn25style V2

A style LoRA trained to reproduce sabotenman's newer illustration style, mainly based on works from around August–September 2025.

V2 was trained on a larger and more carefully selected dataset with multiple characters, outfits, poses, compositions, and backgrounds to reduce character-specific overfitting and improve style consistency.

The main focus of this LoRA is:

- Clean and slightly hand-drawn linework

- Simple cel