V1.0: Trained with Realistic Vision v5.1. Strength of about 1.0 works well. Shorter prompts work better. Try adding in portrait for a closer image.

Side note: if you use SD1.5 it serves as a spooky clown generator. That’s a two-fer!