recommendations for best results :

- 8 steps highly recommended

- use short and clean prompt, avoid long detailed prompt (the model will auto apply lora style)

No trigger word needed

this model will be updated soon to fix identity overfitting