Juggernaut XL for A1111 + Lora + Negative Embedding

Juggernaut XL for A1111 is pretty amazing. I also comes with it's own Juggernaut XL Lora and Juggernaut XL Negative Embedding. Here are sample images and the best settings and sampler to use with this model.
#### Links from the Video ####
Juggernaut XL civitai.com/models/133005/jug...
Juggernaut Cinematic XL LoRA civitai.com/models/120663/jug...
Juggernaut Negative Embedding civitai.com/models/81563/jugg...
#### Join and Support me ####
Buy me a Coffee: www.buymeacoffee.com/oliviotu...
Join my Facebook Group: / theairevolution
Joint my Discord Group: / discord
AI Newsletter: oliviotutorials.podia.com/new...
Support me on Patreon: / sarikas

[@OlivioSarikas]

#### Links from the Video #### Juggernaut XL civitai.com/models/133005/juggernaut-xl Juggernaut Cinematic XL LoRA civitai.com/models/120663/juggernaut-cinematic-xl-lora Juggernaut Negative Embedding civitai.com/models/81563/juggernaut-negative-embedding

[@LouisGedo]

Hi

[@KandooAI]

What an Honor for me that you tried Juggernaut XL :) The Negative Embedding was for 1.5 tho ;) Greetings from Germany ;)

[@TomiTom1234]

Ist diese Model dein Werk? Und für deine Infos, Olivio ist auch aus Deutschland.

[@FusionDeveloper]

I looked at some of the images it has made and it looks AMAZING (the model is downloading while I typed this). Thank you so much for making this!

[@sweetpo1]

Amazing Work! Thank you!

[@KandooAI]

@@TomiTom1234 Ja das Model ist mein Werk :) Ich hab auch schon aufgeschnappt das Olivio aus Deutschland kommt ;)

[@KandooAI]

@@FusionDeveloper Thank you for these nice words :) I´ll hope you have fun with the current version :)

[@FusionDeveloper]

If you didn't know, I found out (by testing) that you can use sub-folders, as long as you have it in the proper folder. for example: \models\Stable-diffusion\SD_XL\ \models\Stable-diffusion\SD_1.5\ and \models\lora\Lora-SDXL\ \models\lora\Lora-SD_1.5\ now I want to go back and organize all my models that way.

[@joskun]

Thank you!

[@Elwaves2925]

Saw this one earlier and was tempted to give it go. I think I will now. Thanks.

[@RazielDark]

I've used this one, it's amazing! i'm glad they made it for SDXL!, and as always, Thanks Olivio!

[@c0dexus]

The negative embedding is for the SD 1.5 version of Juggernaut so it's normal that it doesn't work with Juggernaut XL.

[@BVLVI]

pin

[@freke80]

I thought negative embeddings aren't affected by the model type, since it is a merge of many prompt terms. I could be wrong off course.

[@c0dexus]

​@@freke80 SDXL uses a different CLIP model than 1.5 models

[@AI_opener]

Thanks for the great introduction today! 🤩🤩

[@RodgerE2472]

Great Job as always. Thank you, sir.

[@NotThatOlivia]

Juggernaut XL for ComfyUI is my current favorite SDXL model

[@Dzynerr]

Great video Sir! Please do a video on how to increase image generation speeds on SDXL. Would really appreciate that, thanks!

[@sb6934]

Thanks!

[@twey12345]

Obrigado pelas informações Olivio, mas gostaria de uma sugestão de conteúdo. Você poderia ensinar a fazer animações no A1111 mas somente instalado no PC? Mas sem usar o Google Colab.

[@AgustinCaniglia1992]

The negative embedding is for 1.5 thats probably why it doesnt load on SDXL

[@sergiog377]

Hi Olivio, and ALL, is there an issue with "Embedings" in 1.6 ? I see nothing in the WebUi :( i worked before. Any ideas ? Olivio mentioned it also at 2;53

[@Mat-de2dj]

I've been having problems with the newer version of stable diffusion, the 1.6 one.... for example i cant use hiresfix, it stays in a loop and doesnt give an output. Any idea of what can that be?

[@Avalon19511]

Olivio I wonder if you can help me out, ever since 1.6 I have been getting oversized images after generating an image, I think in one of your videos you showed how to set up the settings, I think after 1.6 something was reset, I'd appreciate the help thanks.

[@cebesten]

Is it possible to create a Lora from juggernaut to have an specific model with the realism of this model? or how would be the best route? just starting in this world

[@christopherthompson1254]

I'd love images of a model wearing a specific (and unchanged by AI) pair of sunglasses, what would be the best workflow to upload this object and then incorporate them in an AI setting?

[@mariotubin]

does sdxl work for mac? because i tried the install, and one of the xl checkpoint can be loaded from the ui drop-down menu

[@WhatDaStat]

3M sampling is also good.

[@joeyc666]

Was that Commander Shepherd at 2:12 ?

[@TigerWoodsLibido]

Can’t get the full version to work with 1111 on my 3060 gpu. I’ll try the smaller 6gb version and see if it works better

[@6Eternal9]

Hi, recently I have been facing an issue from XL models, xl models were working fine but recently I've been getting this error "No operator found for `memory_efficient_attention_forward` ", Does anyone know any fix??

[@joskun]

I need to update my Automatic-1111 to 1.6 Do I do the whole -git upgrade or is there another way of updating Automatic-1111 Also I have the Civitai extension that allows you to create on the site itself yet I could only create with the "Juggernaut Negative Embedding" Do you guys know why it only allows me to do that where as the others shows me the blue download button instead of the create button??

[@harunsivasl4983]

i have a question if you answer i would be appreciated, i was using sd 1.5 automatic 1111 and want to upgrade it to SDXL, if i use juggernoutxl model as base model means i am using SDXL right ? or do i need a automatic111 scratct installation and download sdxl base models and than use sdxl im sorry its a complicated question but i amcomfused a little bit thank you

[@OlivioSarikas]

yes, using the model is what makes it SDXL. but keep in mind that all other things also need to be able to support that. for example you need the controlnet XL models for that

[@harunsivasl4983]

@@OlivioSarikas thank you for your answer i uninstalled everything including brew and re installed everything including automatic 1111 my python version is now Python 3.12.1 its not a problem right ?

[@Arewethereyet69]

Anyway to change presets based on model you use? Hard to remember what to use with what model.

[@user-yj3mf1dk7b]

there are helper extantions like civitai helper. or just make a folders with best imgs and drag&drop to PNG info.

[@Arewethereyet69]

@@user-yj3mf1dk7b, thanks.

[@NTNGChannel]

Hey guys, I have one question. Juggernaut XL or epicRealism, what`s your favorite model for realistic pictures? I tried both and think, they are very similar. Happy New Year to everyone!

[@berke9313]

Its not working. When i try to select it, it just goes back to the previous Checkpoint

[@Radarhacke]

I get much better results with the Chrystal Clear XL Model.

[@Taki7o7]

20% refining is too much imo, no idea why it's set to 0.8 by default ._.

[@workflowinmind]

The TI is not for XL

[@labibmahmud986]

How to install it in mac?

[@dividendenkontor]

😂😂😂

[@azazeal5874]

SpectrumBlendX

[@raydioz]

you need more steps to avoid poor rendering especially w/ loras

[@petitemasque5784]

I praise @KandooAI efforts but hands still look terrible. I will stick with 1.5.

[@KandooAI]

True i can´t deny that ;) But Training is only at roundabout 25-30 % . It will get a lot better in the next couple of weeks/months :)

[@petitemasque5784]

@@KandooAI I hope so, XL has a lot of potential.

[@Larko_-ci3vm]

Can you explain why are you using SDXL models right now, when 1.5 models can do literally anything better?

[@user-yj3mf1dk7b]

i can explain... ppl claiming 1.5 is better are blind.

[@phizc]

If nobody promotes SDXL, nobody will use SDXL, and then nobody will create SDXL models. SD 1.5 is "better" only because people have been creating better models for it the last 12 months. The base SDXL 1.0 model is miles beyond what SD 1.5 was, and derived/descent models have the potential to be incredible.

[@Larko_-ci3vm]

@phizc i completely agree about potential. I'm talking only about the current state of SDXL.

[@Larko_-ci3vm]

@user-yj3mf1dk7b I didn't try SDXL for my own, but from what I can see from Civitai previews, it's clearly worse than 1.5 I would like to be mistaken, so can you explain why you think we are blind about em?

[@phizc]

@@Larko_-ci3vm were you asking the original question to Olivio or people in general? In my case, I'm a sucker for new toys. I experiment to see what strengths it has and what limitations SDXL models still have. Just playing around with it. I also want to try to train a model myself when it starts to get cold. I'd rather heat my room with my GPU than an electric heater 😄. I haven't found any tutorials on how to do that though, just LoRAs, TI, and DreamBooth.

[@fernando749845]

All the eyes are CRAP! How is this better than 1.5? Try again.

[@DrSid42]

It's slow, but at least it makes worse pictures.

[@brunohof2972]

(worst quality, low quality, normal quality, lowres, low details, oversaturated, undersaturated, overexposed, underexposed, grayscale, bw, bad photo, bad photography, bad art:1.4), (watermark, signature, text font, username, error, logo, words, letters, digits, autograph, trademark, name:1.2), (blur, blurry, grainy), morbid, ugly, asymmetrical, mutated malformed, mutilated, poorly lit, bad shadow, draft, cropped, out of frame, cut off, censored, jpeg artifacts, out of focus, gltch, duplicate, (aibrushed, cartoon, anime, semi-realstic, gi, render, blender, digital at, manga, amateur:1.3), (30,30 Game, 3D Game Scene, 3D Character:1.1), (bad hands, bad anatomy, bad body, bad face, bad teeth, bad arms, bad legs, deformities:1.3) negative prompt