lw Lora for SDXL 1.0

1.0 is harder to train I had to try different setting to make it look great.

It's not as good as 0.9 for consistency but when it hits, it hits hard.

I highly recommend to hi.rez-fix the image with ESRGAN-4x with a strength from .25 to .35 that will increase de details and fix the defects in the eyes mostly

Negative prompt seems heavily remove the likeness on the Lora. So keep that in mind.