(Updated 8/14/2023) (Previews made at 512x768 upscaled w/ ImgtoImg)
Now that I'm much better at making Loras, I wanted to give my first one a second go around, now using a higher-res, and bigger dataset (200+ Images used!)
Still based on her Zishy Era,
Trained on SD1.5, so pretty good compatibility with most realistic models
Setting the Lora to 0.7 should do the trick in most cases, a good range to stick in is the 1-0.6 range, with the lowest being 0.6
(If all else refer to Weight Sheet for Guidance)
Trigger Word SabLynn should be all you need, but words like red hair, freckles, can also help if some details get lost
Have Fun! :))))