metadata

library_name: diffusers
pipeline_tag: text-to-image

Juggernaut XL

For business inquires, commercial licensing, custom models, and consultation contact me under [email protected]

Join Juggernaut now on X/Twitter

Run the SFW Version on RunDiffusion

[Prompting Guide Version X (Beginners)](https://storage.googleapis.com/run- diffusion-public-assets/Prompting_Juggernaut_X.pdf)
[Detailed Prompting Guide](https://learn.rundiffusion.com/prompting-guide-for- juggernaut-x/) by Adam (Highly Recommended)

[Juggernaut X on HuggingFace](https://huggingface.co/RunDiffusion/Juggernaut- X-v10)

A big thanks for Version X goes to RunDiffusion (Photo Model) and Adam, who diligently helped me test :) (Leave some love for them ;) )

UPDATE:

As promised, a faster version (Hyper) of Juggernaut for those of you with lower PC hardware. I hope this helps you :) At the same time, I've linked a new [detailed guide](https://storage.googleapis.com/run-diffusion-public- assets/Prompting_Juggernaut_XL_and_Hyper_Guide_v2_By_Adam_Stewarts.pdf) from Adam above. I recommend, especially for Hyper, checking out the guide. It should help you significantly more than my own.

As a community member pointed out to me, ByteDance has announced a Hyper version with higher CFG. As soon as these are released, I will sit down and see if another release is worthwhile :)

The recommended settings can be found a bit further down or under "About This Version". For everything else, as I've already mentioned, I recommend Adam's guide

On to a new round!

With version X, the framework of Juggernaut has been completely rebuilt and captioned with GPT4 Vision. Additionally, the base set has received a completely new structure, which should make it easier for you to prompt certain areas (such as Full Body, Midshot, Portrait, for example).

I'll spare you lengthy speeches about prompting. [The Prompting Guide](https://storage.googleapis.com/run-diffusion-public- assets/Prompting_Juggernaut_X.pdf) should help you more in these areas than a few words in the description ;)

Some areas require multiple attempts to achieve a good result. Especially notable are the "Text","Male Genitalia" and "Hand Poses" areas. All are possible with the model but are still in a "development phase."

For those who have followed my journey in the past, you may remember that I put SD 1.5 on hold after the release of SDXL to focus entirely on SDXL. At the same time, I felt back then that SD 1.5 had already reached its plateau. However, this time I want to take a different approach; I also want to continue providing updates for SDXL for the time being, as there are still improvement opportunities that I want to take advantage of. So, in the coming weeks and months, there will be a few more significant updates :) A Lightning Version will hopefully follow next weekend.

Have a little patience; it's a small adjustment from the old versions. At the beginning, refer to the Prompt Guide (if you encounter any issues), and it should then be much easier for you :)

Recommended Settings Normal Version (VAE is baked in):

Res: 832*1216 (For Portrait, but any SDXL Res will work fine)

Sampler: DPM++ 2M Karras

Steps: 30-40

CFG: 3-7 (less is a bit more realistic)

Negative: Start with no negative, and add afterwards the Stuff you don´t wanna see in that image. I don´t recommend using my Negative Prompt, i simply use it because i am lazy :D

VAE is already Baked In

HiRes: 4xNMKD-Siax_200k with 15 Steps and 0.3 Denoise + 1.5 Upscale

Recommended Setting Hyper Version (VAE is baked in):

Res: 832*1216 (Any SDXL Res will work fine)

Sampler: DPM++ SDE Karras

Steps: 4-6

CFG: 1-2 (recommend 2 for a bit negative prompt affection)
Negative: Only working slightly on CFG 2

HiRes: 4xNMKD-Siax_200k with 3 Steps and 0.35 Denoise + 1.5 Upscale

And now, have fun trying it out. As always, I'm eagerly waiting for your pictures in the Gallery :)

If you liked the model, please leave a Like. In the end, that's what helps me the most as a creator on CivitAI. :)

Last but not least, I'd like to thank a few people without whom Juggernaut XL probably wouldn't have come to fruition:

Dreamlook.AI (Trained 3 Side Sets)

Chillpixel

SilasAI6609

Adam

Model Version: Juggernaut_X_RunDiffusion

About this version:

Recommended Settings:

Res: 832*1216