I trained this LORA on a subset of his work and captioned this subset with waterkuma
.
I selected even smaller subset of artwork that felt closest to zenless zone zero style and captioned it with zezoze
.
I wanted `waterkuma` and `zezoze` tags to have slightly different style, check the test on Corin in the gallery below. In my opinion it's a very subtle difference.
Some of the most popular tags are: twintails, smile, gloves, jacket, jewelry, shorts, socks, shirt, navel, hair ornament, earrings, motor vehicle
Increased uniformity of the dataset.
Changed training settings to eliminate overfitting and added more steps to the training.
Go ahead and upload yours!
Your query returned no results – please try removing some filters or trying a different term.