Photorealistic AI Generated Images

I think I'm going back to school...as in old-school...as in back when I first started getting into this stuff and I was paying for a service, the models were creating almost near outlandish versions of women with nice hips, thighs, breasts that were anime-exaggerated bodies. That's what I liked! But trying to get that with most "realistic" models has always been a challenge for me. When I prompt for such a body I think the models have been trained to "know" those are for older women.

Now I created a workflow that first creates the image in anime style that's trained on outlandish bodies from all the anime/hentai art out there. Then I run it through the model I like using in an image-to-image run, and things come out just fine.

On the subject of creating models and workflows to suit individual tastes: 'He wouldn't be daft enough to create a Checkpoint that solely exists to cover two separate, slightly odd fantasies, would he?



I think he may have just done exactly that...

_2old_ said:

I think he may have just done exactly that...

Click to expand...

LOL Yes he would! Outstanding!

Casshern2 said:

I think I'm going back to school...as in old-school...as in back when I first started getting into this stuff and I was paying for a service, the models were creating almost near outlandish versions of women with nice hips, thighs, breasts that were anime-exaggerated bodies. That's what I liked! But trying to get that with most "realistic" models has always been a challenge for me. When I prompt for such a body I think the models have been trained to "know" those are for older women.

Now I created a workflow that first creates the image in anime style that's trained on outlandish bodies from all the anime/hentai art out there. Then I run it through the model I like using in an image-to-image run, and things come out just fine.

View attachment 3850451

Click to expand...

Here are some examples of what I'm talking about.

First model that generates the initial image at 480p: Fantasm | Pony XL
Second model used for the image-to-image at 1080p (after an upscale to 576p): GonzaLomo XL/Flux/Pony v4.0 Unity XL DMD

The Pony model creates some fantastic hips, waists, and breasts that the XL model later sticks to so long as keep that CFG value around 0.56 or so. The drawback so far, but I'm hoping that's just because of my lazy prompting (I don't describe the background and surrounding too much at all) is that backgrounds don't look all too realistic, BUT, I'm not in it for the scenery. You'll also see the detail and especially the faces are generally a disaster, but the GonzaLomo model does a great job of just keeping the inference and doing its thing to clean them up.