Model made of Alina Nikitina - All images used in training are after her 18yo brithday
Description
FAQ
Comments (5)
What SDXL checkpoint do you recommend to use?
Cyberealistic, also I use adetailer
Only 11 images with 512x768 resolution for SD1.5. I can generate the face but they all look the same.
this lora is for SDXL, also try a lower weight like 0.7
This is just an hint. Using lower res images on a higher res model will result in strange deformations. I got long necks in some images when generating on SDXL resolutions. You get better images if you train on XL resolution. 512x768 is SD1.5 which I have used for training before.