Anna Vlasova (1999 - ) is a Russian Instagram model. This LoRA was trained on 57 images of her face from around age 19 or 20 to the present.
Try starting with a weight of 0.3 on this LoRA and 1.3 on the trigger word, (anna_vlasova: 1.3). If you don't want closeups of her face, try describing her body and clothing from head to toe in your prompt and putting the trigger word after her bodily description.
Description
FAQ
Comments (4)
I can't get this to look like her, though the results are still nice. Which checkpoint do you recommend?
The sample pics were all using Objective Reality 2.0. I seem to get the best results using that one as well as the original SD 1.5 and Realistic Vision (5.1 at the time.)
@baddrudge Thank you--I'll try those out.
Finally got back to re-testing this. I typically stick with Cyberrealistic and it was the model strength settings--0.7 seems to work much better.