Hello, I present you my first ever LoRa model. Model is trained on a dataset of 83 images in 1600 steps. This is it's 7th version. I endeavoured to make it as accurate as possible, comfortable to use, flexible and without major errors or big compromises. For demonstration purposes I did not use detailers nor inpaint (even if I'd love to correct eyes on some), so you should be able to get the exact same results. Check out my metadata for reference.

For more comfortable use, please follow these instructions:

Trigger words:

Character: Suraya_by_Z666Ful

Standard clothes: sweater_1, jeans_1

Optional tags (positive/negative):

ponytail

gradient hair - If positive, lot of blue. If negative, almost no blue at all. If neither, something in between.

inverted nipples - Keep them in negative unless you want them. If they are neither in pos. or neg. it tends to do some kind of mix.

sweater lift - Does not happen often, but in some cases it lifts her sweater without being prompted. Negative might help.

topless/bottomless - if only sweater/jeans are wanted

I hope it works as well for you as it works for me.

Have fun and consider checking out the original artist.