Hugging Face's logo

TheImposterImposters
/
URPMPonyXL-HybridV1

Text-to-Image
Diffusers
Safetensors
StableDiffusionXLPipeline
Not-For-All-Audiences
Inference Endpoints
Model card Files Community

Uber Realistic Porn Merge 🦄PonyXL-Hybrid (URPM)|✅XL & Pony LoRAs |✅ControlNet - PonyXL-Hybrid V1

  • ID: 790652
  • Author: saftle
  • Type: Checkpoint
  • Base Model: Pony
  • Base Model Type: Standard
  • Model Version ID: 923681

Daily Release Campaign has ended!

See here for more info: URPM is back - Daily Releases incoming!

💖Join the Patreonif you want to support our work & even earlier access to models being beta or alpha tested!
🗣️Join the URPM Discordfor Updates or Support!

PonyXL Hybrid is compatible with SDXL LoRAS, Pony LoRAs as well as ControlNet!

If you find any incompatibilities, please let me know and I will try and improve this in future versions. I really wanted to make a model that no longer required you to choose SDXL for specific things, or Pony for other things.

My recommended Hybrid-PONY prompt scores and negatives:

Prompt: score_9, score_8_up, score_7_up
Negative: score_1, score_2, score_3, score_4, text, watermark, branding

My recommended SDXL-Only negatives:

Negative: text, watermark, branding  

Additional Pony Specific Prompt Help:

First off, this requires a bit of understanding as to how Pony works. I recommend some light reading first if you are new to Pony: https://civitai.com/articles/7052/pony-prompts-for-dummies

Additional General Prompt Help:

Be sure to check all of my sample images. Every image was created without TIs, LoRAs, adetailer, facefix, upscaling, hires fix, etc. They are completely raw outputs. You can drag them into PNGInfo, if you're using Automatic1111, Forge, etc.

Be sure to also try a small prompt with very little negatives or TIs/LoRAs first before adding anything unnecessary. Not all models work with invasive FIX-ALL-THE-THINGS negatives. Sometimes the model can do it without them. Also, a little goes a long way. You don't need to spam every keyword like we used to have to do with SD 1.5, in the hopes that we'll get something close to what we want. 

People I need to thank:

  • LJLeb for creating the amazing sd-mecha/comfy-mecha. Without these, I wouldn't have been able to merge Pony and SDXL so well together. Secret: Use Train Difference and Clamping.
  • My Patrons and followers for having waited patiently for 2 years for me to come back to the scene and were still willing to support me. You guys changed my life 2 years ago, and I owe you all everything.
  • My Team for sticking with me these 2 years. We've done a lot together, and I can't wait to see what we do next.
  • S1LV3RC01N for always being on the lookout for new techniques and methodologies that I can learn from. Thanks to you, I learned about Train Difference.
  • 6tz for finding a few issues in my Pony model, that I was able to solve at the last minute
  • CIVITAI for providing this community and effectively changing the trajectory of my life. I owe this platform everything as well.

Creators I need to thank for my SDXL/Pony-Hybrid models:

Some models may have once been included in the merge and are no longer, or I used them as unet block references.

Liability:

In no event shall I or my team be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the use of this model. Please render responsibly.

First Pony/SDXL Hybrid!
- It can use scores unlike some Pony/SDXL hybrids making it more of a Pony model without losing too much of its Pony roots.
- It can use Pony LoRas as well as SDXL LoRAs, and even ControlNet models made for SDXL

Also it by default is incredibly realistic without having to prompt to make it realistic. I hope you all enjoy! This was a labor of love!

Content copied from Civitai: Uber Realistic Porn Merge 🦄PonyXL-Hybrid (URPM)|✅XL & Pony LoRAs |✅ControlNet on 2024-10-23 09:06:18.059996+00:00

Downloads last month
29
Inference Examples
Text-to-Image
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.