Generates a likeness of the actress, Natalie Portman.
Likes a guidance value between 4 and 6.
v1.2 is trained at 512px with the trigger word: p0rtm4n - using a lower learning rate to vastly improve glitches and frozen faces
v1.1 is trained at 1024px with a trigger word: p0rtm4n
v1 is trained on 512px
Description
FAQ
Comments (4)
You make great loras! You use diffusion-pipe? What is the learning rate you are using now?
Have you settled on 512px @lower learning rate as a better approach than 1024px?
Would you mind sharing your settings? Or at least the resolution/learning rates that you've experimented with?
547687Feb 12, 2025· 1 reaction
CivitAI
Maybe she's born with it.. ;)
bro i saw u change resolution from 1024*1024 to 512*512, could u explain it bro