This LoRa was trained with 75 images, which were cleaned of text.

All examples were processed using only Face Adetailer.