Join The Tinkerer on Whop to download this model and many more — all for one monthly payment. Plus, get early releases, private tools and a lot more. 👉 Join on Whop
CyberRealistic Pony blends all the charm of Pony Diffusion with the striking realism of CyberRealistic. The vibe? You get everything from adorable to bold (sometimes both at once) with crazy-detailed textures, moody cinematic lighting, and a hint of AI flair.
🧠 How to Use It
Sampling method: DPM++ SDE Karras / DPM++ 2M Karras / Euler a
Sampling steps: 30+ Steps
Resolution: 896x1152 / 832x1216
CFG: 5
Clip Skip: 2✅ Recommended Prompts
score_9, score_8_up, score_7_up, (SUBJECT), Negative Prompt examples:
score_6, score_5, score_4, (worst quality:1.2), (low quality:1.2), (normal quality:1.2), lowres, bad anatomy, bad hands, signature, watermarks, ugly, imperfect eyes, skewed eyes, unnatural face, unnatural body, error, extra limb, missing limbsit's also possible to remove the normal pony tags:
(SUBJECT), Negative Prompt examples:
(worst quality:1.2), (low quality:1.2), (normal quality:1.2), lowres, bad anatomy, bad hands, signature, watermarks, ugly, imperfect eyes, skewed eyes, unnatural face, unnatural body, error, extra limb, missing limbs🛠 ADetailer Settings
Adetailer model: face_yolov9c.pt
If you only want the main face being refined set 'Mask only the top k largest' to 1.☕ Support
Enjoying the ride? [Buy me a coffee] – but only if the Pony delivered something you actually like.
⚠️ Disclaimer
This model might generate sensitive content. Whatever you make with it is on you. Don’t do anything weird and try to blame the horse.
🔗 Links
Backup location: huggingface
💡 Need Better Prompts?
This custom ChatGPT was made to whip up top-tier prompts just for this model:
🔗 [Try it now on ChatGPT]
Description
Custom VAE included in this version.
Enjoy!
FAQ
Comments (34)
Thank you for creating this model!
LOOK~WE HAVE THE SAME RING~:)
Twinsies! ;)
ring party
@Dever You should never say that in a Pony model thread, you might veer the discussion off into a tangent you weren't ready for :)
@Sharinel seriously? LOL~I dont have idea about that~
@Sharinel @BIG_A Ah, the unspoiled youth.
awesome checkpoint, loving it so far! getting a lot of cross-eyed faces - any pointers on how to avoid them?
Crosseyed in negative? Didn't notice this problem
Do you have this problem in close-up face images? I will test it tomorrow.
It was really complex to create this checkpoint (have a lot of beta versions) - maybe I can change it or see where it went wrong. But again didn't notice this problem
@Cyberdelia no, more like in full body shots. I tried cross-eyed in negative, that helped actually! would have liked to post examples, but I have the habit of deleting misgens right away
Cyberdelia, you note for version 2.0 that "the choice between the versions depends on what you aim to achieve."
Could you share a few words on what version 2.0 is stronger at than version 1.0 and vice versa? I just downloaded 1.0 yesterday and think it's great. I will gladly try out 2.0 too.
Euhm...how can is say this tactical... Version 2 is less likely to cause you problems than version 1 ;)
@Cyberdelia Can you be a little less tactical?
@Cyberdelia Obviously there is no technical issue, no 3 hands or anything - which means that it is probably a NSFW/SFW issue. One probably more NSFW than the other. Luckily acting like a bashful 12 year old and not telling us which is which means we have to download both to compare. A cunning way of getting double downloads :)
@direct22202 Version 1 often generates characters that appear younger by default. This issue is somewhat resolved in version 2. In comments on other checkpoints, I've seen that this is a common problem, but it's something that Pony does.
@Sharinel I apologize for any confusion. Seeing other comments, I realize my response wasn't as clear as I intended.
Great model for what it's trying to do, but I find that all caucasian female faces look very similar. Even looking through the sample images, I see a lot of familiar faces.
Checkpoints are trained on millions of different faces all called : "girl" or "woman".
So, of course all checkpoints tend to produce a similar face. That's what prompts are for. DESCRIBE what you want, and you'll get something different. Either by prompting races or facial features. Or use LoRAs. The "problem" you describe is a perfectly common thing to all SD models.
Beautiful model!
Great model, thanks! I did compare v1 and v2 and I much prefer v1, it "feels more Pony".
I have to test it further but very nice model! A hyper version would be appreciated. Have a great day.
Hands down the very best realistic Pony around. Outstanding results I get out of it. In one word : wow. I mean it.
Thanks!
P.S: v1.0 rocks !
Interesting.
Good model, but it seems to lean heavily towards "overly contrasty" lighting (this can even be seen in the example pics).
Is this an intentional aesthetic, or something you are looking/open to change?
The custom VAE is doing this (it's intentional :) ) But I can change that. Maybe I will change the V2 and V1 and upload it to huggingface or on Civitai.
@roe2 please check https://civitai.com/images/12531903
@Cyberdelia Hi Delia! I thought my ComfyUI is broken 🤣 I'm having the same issue/feature(?), first KSampler color is perfect, but after the second KSampler it went and oversaturate things. Also facedetailer and hand detailer will add even more contrast. Thanks for the info I will try to use other VAE. So far I'm having fun with the model as always, keep them coming!
@dkorbat I've got alternative editions of V1 & V2. Currently busy with sample images, planning to share them on Civitai soon. Fingers crossed folks won't mind the uploads.
@roe2 & @dkorbat new versions online. Hopefully this works better for you
@Cyberdelia Thank you. I've found I prefer the "a" versions after noticing the saturation difference when first tinkering with 1.0 and 1.0a.
@markovochka280 good to hear!
@Cyberdelia Yo thank you for the quick turnaround! I'll give this a shot and report back :)
@Cyberdelia Thanks for taking the time to update them!