This is fucktastic. A porn-generating machine for Pony/PDXL.
It's focus is on creating high quality, photorealistic, but exaggerated images with a focus on women.
Also available on mage.space.
Do not use the Normal scheduler. Use Simple or Beta!
Use euler_a, dpmpp_2m_sde or dpmpp_3m_sde as your sampler.
To use the faster low-step settings, you need to use the DMD2 LoRA. This allows you to use 8 or 16 steps and generate very fast. Remember to generate in 2 passes, either using "Hires fix" in Forge/A1111, or by creating and chaining a second sampler node in ComfyUI. A sample workflow if included in the sample images of the model. Use this as a reference.
Each sample/preview image contains the used workflow. Here's a quick article with simpler more beginner-friendly workflow. This is a recommended starting point.
Recommended Settings
Steps: 20
CFG-scale: 7
Sampler: Euler A
Scheduler: Beta
Description
FAQ
Comments (16)
Such a great evolution of this model over time. Always fun and surprisingly versatile for all occasions - still one of my favorites!
Thanks for the kind words!
It's one I'm always tinkering with in the background. The further away I can get from "Ponyface" the better, has always been the goal. Obviously at reduced pony knowledge and anime style for a realistic model, but still very potent and capable!
absolutely criminal how underrated this checkpoint is. this one stays in my rotation, incredibly good model
Superior model. v 5.2 is truly fantastic. Thanks again for improving an already excellent checkpoint (5.1).
Seriously lacking in racial/ethnic diversity. While the quality of what this outputs is good, a VERY simple prompt "a black man" consistently generates a WHITE man. I thought something was wrong with a more complex prompt I was using when one black male character kept getting 'replaced' with a white man in the image generated, until I just ran the simplest possible: "A black man" and ran it several times. Each time yielded a picture of white man. It's like the checkpoint doesn't even 'understand' the concept of a black person. Not cool. I'll have to look for another checkpoint that uses a more diverse training set.
Thanks for the feedback. I'll try to improve this in a future version.
Sometimes concepts like this gets removed during version merging. But I should be able to restore it for sure.
Did you tried to use the more logical term african/east nigerian man ? words with different meanings can indeed be hard to understand by AI , i often try countries or ethnic group names instead of political terms.
"ebony" tends to work best for me
I mean, you can't complain about morality when you're doing something that isn't exactly moral in the first place, wtf cares about diversity, not everything needs to be diverse anyway, no one goes to a black church and demands diversity...
Woke moaning aside, its good to have diversity in the models, sometimes one may want to see other type of men banging the girls. Simple as that. Wokes, kiss my ass.
It's hard to ensure that all biases are good at all times. I'll add it to my list for the model.
I don't know what I'm doing wrong but I get really burned images with this model. I have to go as low as 1 CFG to get bad quality realistic pics.
just pick one of the random example image of the model. move it to the comfyui. it will load the workflow... use comfymanager to download everywhing you need and try again... it should redo the same picture you used as example. after that just try different prompts
@mdtaibo82207 This is great advice. I wish it was more clear to beginners on Civit how this works. And by this I mean, if a "beginner" user (i.e. someone that hasn't confirmed that they know this), looks at a ComfyUI-sourced image, it could pop up a tutorial about how to do this.
@6tZ you could use gemini to learn too... just lie about the nudity, or theme of the image and the name of the checkpoint... for example with this one. I have this XXXX problem when trying to do a picture in (this one its based on sdxl) sdxl. help me. it should solve the issue eventually xD
Amazing



















