Improved character consistency and facial accuracy.

Better anatomy and lighting handling.

More stable results across different poses and scenes.

- Trained on 146 images

- Improved body proportions and expressions

- Works well for portraits and action scenes