v0.7
Original 44 images (res fixed for SD XL), blip captions +
a bunch of Summer images, WD14 captions
(all captions carefully manually edited)
Actually I'm using the Prodigy optimizer with settings similar to those of civitai.notion.site SDXL training overview
v0.4 with 44 images (original res variable from 1509x2028 to 900x1280, all images are manually captioned)
v0.3 with 30 images (all manually captioned)
Description
from my first tests on SD 1.5 emaonly this LoRA version generates really hilarious images XD
FAQ
Comments (5)
I love the concept and I'm looking forward to the next version!
do the v0.2 version is already based on the larger set of images?
ps: why I dont see the prompts in your gallery?
for the style, doesnt need the trigger world.
Your covers are really getting better day by day!
(i should say: "training by training"!)
Do you have some of this complete comics.
Are you able to read the stories in their original language?
Vuoi unirti a un piccolo gruppetto telegram di utenti italiani di civitai? li sto raccogliendo man mano ne trovo