사용자 LoRA
<lora:shoshikastyleIllustriousXL:1>,
No keywords required, trained on 42 images. Should be compatible with character LORAs, but might need to fiddle around with the weights.