This LoRA model is trained to reproduce the facial features of an original character named “Kasuga Tsumugi”.

The model focuses primarily on facial identity, including face shape, eyes, nose, mouth, and overall facial balance.

It is optimized for realistic, photo-style portrait generation, especially close-up and bust-up compositions.

This LoRA is best used for generating consistent facial appearances of the same character across different poses, expressions, and lighting conditions.

It performs well with front-facing and three-quarter view portraits, and is designed to maintain facial consistency rather than clothing or background details.

This model is not intended for stylized, anime, or heavily artistic illustrations, nor for generating exaggerated expressions or extreme poses.

For best results, use realistic photography prompts and avoid strong style modifiers.