Name: Photonic Pony Fusion
Author: StecFX

Photonic Pony Fusion

Welcome to Photonic Pony Fusion V2!


 Building off of V1 I have continued to refine and improve the checkpoint and to iron out some of the quirks of the original. It does male anatomy more consistently now and overall big improvement in realism.



<ul>
 <li>
 
 On A1111/ForgeUI I recommend DPM++ 3M SDE + Karras
 
 </li>
 <li>
 
 Steps 30-40
 
 </li>
 <li>
 
 CFG: 3-6
 
 </li>
 <li>
 
 768 x 1024 or 768 x 1100 Upscaled 1.5 - 2x
 
 </li>
 <li>
 
 Hires Fix, 10-15 Steps, 0.30-0.40 Denoising, 1x-ITF-SkinDiffDetail-Lite-v1 is the best on A1111 and ForgeUI
 
 </li>
 <li>
 
 I suggest using ADetailer both for face face_
 <a href="http://yolov8n.pt" rel="ugc noopener noreferrer nofollow" target="_blank">
 
 yolov8n.pt
 
 </a>
 and eyes mediapipe_face_mesh_eyes_only with appropriate prompting for full body images of people but it's not really necessary for close up shots.
 
 </li>
</ul>



 
 Also I don't believe in locking my models that are on here or available on TensorArt or CivitAI behind a paywall like some do but I gladly accept donated --&gt;
 
 <a href="https://ko-fi.com/stecfx" rel="ugc noopener noreferrer nofollow" target="_blank">
 
 
 Coffee and Beer
 
 
 </a>
 
 &lt;-- your donations help me to keep motivated to continue making new checkpoints and LoRA's for all our enjoyment! To those that are generous enough to throw me a few dollars I thank you! &lt;3

Photonic Pony Fusion is my first Pony checkpoint model with a focus on "Realism" and of course with an added male twist.
 



<ul>
 <li>
 
 On A1111/ForgeUI for realism I recommend Euler a with Automatic, (DPM++ SDE + Karras is also good),
 
 </li>
 <li>
 
 Steps 25-40
 
 </li>
 <li>
 
 CFG: 5-7
 
 </li>
 <li>
 
 Size: 512x768 or 768x512
 
 </li>
 <li>
 
 Hires: Denoising Strength: 0.30-0.50, Hires upscale: x2, Hires steps: 15-20, Hires Upscaler I highly recommend: 1x-ITF-SkinDiffDetail-Lite-v1 (R-ESRGAN 4x is also good).
 
 </li>
 <li>
 
 I rarely ever use ADetailer on this model but it could also help bring about some extra realism, which I have seen in a few test images but can also sometimes bring you to the uncanny valley if using Adetailer I suggest you have your settings to save the before and after image to be safe.
 
 </li>
</ul>

 
 Also I don't believe in locking my models that are on here or available on TensorArt or CivitAI behind a paywall like some do but I gladly accept donated --&gt;
 
 <a href="https://ko-fi.com/stecfx" rel="ugc" target="_blank">
 
 
 Coffee and Beer
 
 
 </a>
 
 &lt;-- your donations help me to keep motivated to continue making new checkpoints and LoRA's for all our enjoyment! To those that are generous enough to throw me a few dollars I thank you! &lt;3