V1 does heavily creates watermarks and other bullshit
V2 trained without watermarks on bigger dataset
Support the original artist:
https://x.com/LollipopconArt
Description
First release
FAQ
Comments (1)
nice lora captures the style well but damn the watermark is strong