Another Pony from Gonzales , to create mostly realistic Pictures. Again as DMD Version for fast Speed and good results.
Have fun with Gonzales-NSFW-PonyV1/V2
Best Performance for me on Forge is
Sampler: LCM or LCM Karras
Steps: 9-11
cfg-Scale: 1-1.3
The animation are created on images made with this model on mage.space using img2vid.
Gonzales NSFW Pony V3:
as before but use cfg 1.5 -> allows using little influence of negative prompt with better results
Description
Version 2
Little Variations - hope you like it.
Same settings as Version 1
Sampling LCM / LCM Karras
Steps 10
Cfg 1
Again a very realistic model
FAQ
Comments (18)
This model is INCREDIBLE!
crazy mix. how did you merge dmd into this model? the detail is unrivaled imho
Looks good I'll have to give it a try.
How good is it at doing different ethnicities?
Most Pony models have issues.
posted creation - no problems - try yourself - african woman next chinese man e.g.
I can't draw a normal dragon. Tell me what I'm doing wrong?
hi foxilisik271 - maybe depending on your prompt and what kind of dragon do you want. maybe send me your prompt? in the meanwhile i will have a look if i have the same issues.
@Gonzales007 Thank you very much! I think you also noticed that the dragons in your model have problems with their limbs, especially with the hind ones (including those you published) they tend to be anthropomorphic (human-like).
man this good
this crazys model,, WTH broo greaaat work
uhhh wtf? this is insanely good
uhh, what a crazy good model. It’s unbelievable. I’m already hyped for the next one!
Great pony model, follow the prompt very well, very realistic and faithfully reproduces subjects of various ages. Fast with excellent resolution. Thanks for sharing it.
Mind blown. The best Pony model I have used so far.
Congrats! This checkpoint is truly great! Most realistic pony checkpoint I've seen so far. Do you have plans a third one soon? thanks!
Garbage, doesn't obey prompts
Thats due to DMD2 being built in. It allows for really faster rendering speeds but the drawback is that it requires a very low CFG and no negative prompt. And it can be run on way less VRAM. So its just a tradeoff you gotta make.
How do I train lora on DMD models? usually comes out blurry when I train it on these kinds. Awesome checkpoint btw.
Best model I used. Insanely good. Thank you!
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.