So, I have been saying for a while that if I get 500 Likes, 100 Followers or 10K downloads, I will post the more cartoonish model that I used for my example images...
So here it is, well, mostly, I renamed it and made it FP16 and turned it into a SafeTensors.
So, let me apologize now for the horrible model that I am about to unleash onto the world!
Trust me when I say, it's not that great, but it does work... I have been using it for half the example images for my LoRA's after all...
You have been warned!
Though, if anyone gets anything out of this, post some images, give it some likes, maybe some buzz, but that I don't care that much about and I'll just use it to tip other people... but yeah...
Also, like my LoRA's, the default keywords I use to make the example images are as such:
Positive:
(best quality, masterpiece:1.2), photorealistic, ultra high res, intricate detail, Exquisite details and textures
Negative:
cartoon, anime, manga, comic, animation, 3D, (yellow, gold, grayscale:1.5), (monochrome:1.5), Paintings, sketches, (worst quality, low quality, normal quality:2), lowres, (6 more fingers on one hand)
Also, I use Euler a at 20 steps and a CFG of 7, HiRes Fix is set to Latent with a De-noise of 0.7, I use the Extra Noise Modifier set to 0.6 and I use GFPGAN at 1...
Description
Positive:
(best quality, masterpiece:1.2), photorealistic, ultra high res, intricate detail, Exquisite details and textures
Negative:
cartoon, anime, manga, comic, animation, 3D, (yellow, gold, grayscale:1.5), (monochrome:1.5), Paintings, sketches, (worst quality, low quality, normal quality:2), lowres, (6 more fingers on one hand)
FAQ
Comments (5)
I really like your Loras, so what better to try them than your checkpoint? :-) I don't like many those cartoon/cartoonish ones around, by I like the cartoonish style of yours.
Are you planing to release the one more realistic that you use? I would have to try it, at least. Again, the mix of both your loras and that checkpoint looks great.
Take your rest, but return latter all powered up for more!
Thanks for the kind words!
If you look at any of the newer LoRA's marked [PrnStr], at the bottom of the description has a section marked "Example Images" that will tell you where the more realistic model can be found.
Found it, thanks. Did you found any significant improvement between v6 and v7?
@WildhawkPT To be honest, I went back to V6... maybe it's because I didn't play with 7 enough, but I think 6 is just a little better at a lot more...
@Meowzorz Understood, thanks for the enlightenment. I'm still waiting for the repo approval, then I'll start exploring it with your lovely girls and posting the results.