The Lora was trained with two datasets:
1) realistic images (photographs), tag: preggo-r*
2) anime images (drawings), tag: preggo-a
Use in positive/negative as you will, based on the result you're aiming for
*If going for "preggo-r", but not liking the photorealism (Pony is terrible at it) put "3d" in the negative to dull the effect
