saba

Saba is a character/person LoRA trained to reproduce the face, hair, and overall look of Saba.

This model is intended mainly for realistic portraits and half-body images.

Trigger word: saba

Available versions are provided for different base models, including SDXL and Flux variants. For best results, use the matching base model listed on each version page.

Recommended LoRA strength: start around 0.7 to 0.9 and adjust as needed.

Best at:

- close-up portraits

- natural skin detail

- clean lighting

- realistic compositions

Less reliable at:

- extreme poses

- very crowded scenes

- heavy stylization