Weighting up to 1.00 gives a more realistic image. Up to 1.5 will make a more abstract image.
Using drawing or illustration in the prompt will create better images.
A companion lora called Visage has been built on the same foundation using portraits of females in the dataset.
Comments