This model was trained for realistic portrait generation based on original reference photos.
Works best for close-up or half-body shots under cinematic or natural lighting.
Use moderate CFG (5–6) and 30–40 steps for optimal detail.
Base: CyberRealistic Pony v14
Clip Skip: 2
Description
First training run on 600 images, focused on realistic likeness and neutral lighting.
Best used for close-up portraits (512–768 resolution).