I pretty much created the model with my usual way using fluxgym, except I used more images (79 rather than 20-30) than usual and I made the size of all images, per recommendation to 1024 x 1024 at 150 DPI. This size required me to upscale many of the images I used. I gathered them from many of the digital comics I own since I was a member of the Jab site a while back. Most of the images where of woman in varies poses and backgrounds. Many had multiple, characters in the picture, often two. I wondered if this was a “no no” or not, but so far doesn’t seem to be an issue. I’m not getting multiple characters unless I prompt for it. I made sure I also had a number of pics of a single a person.
I also included some men, boys, girls.
Also because of online recommendations I removed most all word bubbles. Yeah this took allot of time because you know, comic book source images! I couldn't use Photoshop Ai to get rid of most of them because Photoshop #@&#!~ expletive, restrictions! I used Photopea to remove most of the word bubbles and some hand digital painting in for minor fixes.
Because this is Jab, many of the subjects in the images where in skimpy clothing or none at all. As a matter of fact it was hard to find images of the woman in dresses. I did find a few. I used AI in Photopea to add dresses to a couple of the images making sure they matched with the style.
Still I only had 5 or six images where the female subject was fully clothed.
This too doesn’t seem to be too much an issue so far. Looks like I can prompt clothes on the characters just fine.
Also being Jab, I included a few explicit images as well.
