Finally OUT! ControlNet SDXL within A1111 🔥

[@Paulo-ut1li]

Much better than ComfUI! Those noodles and custom workflows with broken custom nodes are just crazy to manage.

[@hmmrm]

ive been searching for a simple explination and i found yours is the best , thank you

[@vidivy]

You are bringing so much value!!! Thank you very very much!!! ❤🎉

[@frankwirthonlinede]

great instructions. so much relevant information in such a short time. and then with subtitles. great. thank you

[@cpworring]

7:00 - If you're having problems with the bot model anatomy "bleeding through" in the generated image, you can change the model in PoseMyArt by clicking on it, then on "Change Model" (the T-pose icon in the lower left corner) and then select a more realistic human. It will keep the pose of the old model, so if you chose a premade scene with bot models, you can easily change them to more realistic human models.

[@ChristerHolmMusic]

Excellent! I will go get those models now :)

[@KaptonTimo]

Tysm for the detailed guide!

Clear and accurate as always! Thank you

[@aaronsj80]

Awesome, I've been waiting for Control Net to get an update for SDXL.

[@ac3d657]

Its been out for ages.

[@aaronsj80]

@@ac3d657 Last time I checked, the ControlNet extension to Automatic1111 hadn't been updated to used SDXL yet. But it has been quite a while since I checked.

[@LaCarnevali]

yeah but not many people know - that's why the video!

[@happyme7055]

Hi.. thanx for that very instructive tutorial - is there a SDXL version of lineart out there?

[@LaCarnevali]

huggingface.co/TencentARC/t2i-adapter-lineart-sdxl-1.0

[@marcus_ohreallyus]

Thanks for this. I decided to give SDXL a shot after seeing that I can use controlnet with it. I used the JuggernautXl model, which is supposed to be one of the best. I just don't like the output at all because IMHO it just doesn't match the look of sd1.5. Maybe I don't know how to control it yet, but every image looks overly stylized.

[@LaCarnevali]

It's not amazing yet - I think 1.5 is still better re conditioning. Try changing the controlnet weight, and try different models.

[@nadk666]

Thank you so much! I was searching for this exactly. Any idea if SDXL has any controlnet model for face expression which works well? Maybe an updated version of Mediapipe for SDXL?

[@Satscape]

SDXL still seems to be in it's early stages. It's exciting where this is all heading!

[@ac3d657]

Lurk more

[@Satscape]

When people tell me what to do, I usually do the opposite. If Lurking is a popular thing to do, then my own channel must be really popular because no one posts comments. 😁✌

[@w000w00t]

Pay no mind to the trolls, i share your excitement about SDXL, I think everyone is still jaded from the 2.0 release lol

[@Satscape]

@@w000w00t 🤣 Yes indeed!

[@estebanmoraga3126]

Hello! A somewhat related question, how do I get SD1.5 instead of SDXL if I so choose?

[@ExspiravitMundi]

This models of ControlNet works with models from 1.5? Thanks

[@LaCarnevali]

For SD 1.5 models, you need to use these: huggingface.co/lllyasviel/ControlNet-v1-1/tree/main

[@ExspiravitMundi]

Thanks, Super Fan of you.@@LaCarnevali

[@juedbe]

Laura! Love your content. Thank you for all the work you put out. Been having problems running some Preprocessors in ControlNet and hoping you can help me. I've tried updating, deleting and reinstalling, I've tried running SD a1111 in different browsers, redownloading the models, tried etc., and I just keep getting the following error: " urllib.error.URLError: " Anything come to mind that could help me solve this? Like I mentioned, only a few preprocessors work. For example, for "Depth," only the "depth_zoe" works, all others give the error. For "OpenPose", none of the preprocessors work. This happens both in SDXL and in 1.5 models. "Canny" on the other hand, they all work. Any ideas would be great. Thank you in advance.

[@Dannysingh-r7b]

Using chair and sofa images, I have finetuned a stable diffusion model. However, the model does not generate a living room using the chair and sofa. How to train model on multiple subject? Please help me !!!!

[@antor44]

Hi Laura! Thanks for another informative video on ControlNet for Stable Diffusion XL. I'm really curious to know if Inpaint is compatible with the XL version and if it's effective for editing hands and faces in XL. Would you consider making a video tutorial on using Inpaint specifically for those areas in the XL version? It would be incredibly helpful! 😊

[@LaCarnevali]

yep, thank you :D

[@user-pc7ef5sb6x]

Is there SDXL controlnet outpainting?

[@artificialbeat]

Hello Laura you can make a video for open pose? or tell me what openpose file i need to download, we love you

[@iqluevano]

Hi, I know this message has nothing to do with this video specifically, but do you know anything about if there is a solution for SD on macOS Sonoma, your installation video worked perfectly, until yesterday when it updated :/. Thank you very much, I love your videos.

[@LaCarnevali]

SD is not great for Mac as it needs NVIDIA GPU (not supported on Mac). You should try Colab Pro (cheaper option, but sometimes annoying as loosing connection), or Diffusion Hub/Think Diffusion

[@raviswami9852]

SDXL doesn't work with Automatic 1111 when running in Goggle Colab and I get the errors you mention - is this to do with the ControlNet XL models since currently I'm using the old models...?

[@LaCarnevali]

Yes, you cannot use the old controlnet models with SDXL. Would recommend comfyUI (free), or Diffusion Hub / Think Diffusion

[@BetterHumanz]

hmm no XL Tiles

[@ac3d657]

hmM nO xL tIlEs

[@LaCarnevali]

prob need to wait a bit

[@mr.entezaee]

How long do you think it will take to add all the models?

[@LaCarnevali]

5 minutes

[@mr.entezaee]

@@LaCarnevali No, I mean other models like Lineart

[@mr.entezaee]

Inpaint or Reference

[@neehime6]

Hi. Do you have a guide about how to install SD in pcs with amd graphic cards? Love your vids as always. Best regards!

[@Underp4ntz_Gaming_Channel]

sadly openpose doesnt work and also not with pony :(

[@LaCarnevali]

Have you tried xinsir models? They are currently the most efficient for SDXL. kzbin.info/www/bejne/anaqlKSimMSnosksi=jOBp_AhtgLswz1re

[@2thecurve]

I need a XL tile

[@karmitfrog7968]

oh finally someone said comfy not user friendly

[@dreamzdziner8484]

Nowadays having comfy has become like a status symbo🙄. But I never found such node based tools enjoyable. Too much complication.

[@LouisGedo]

👋

[@CharisTsevis]

Ciao Laura, Avrei bisogno di nuovo del tuo aiuto. Ce l'hai fatta di correre SDXL Control Net su mac? Io ricevo sempre il message che "Python crashed".

[@LaCarnevali]

Ciao! SDXL non va sul mac purtroppo. Devi provare Colab, o ThinkDiffusion/DiffusionHub

[@CharisTsevis]

Grazie tante@@LaCarnevali Sempre d'aiuto. I miei complimenti di nuovo per il tuo canale.

[@-Belshazzar-]

wow the openpose in sdxl is really really bad compared to 1.5, honestly, 1.5 is still way better than sdxl in almost every way except prompt accuracy

[@-Belshazzar-]

it actually doesn't work at all. using the t2i-adapter_diffusers_xl_openpose do practically nothing, so I am not sure why it's even on the download page

[@-Belshazzar-]

Ok I found out why, for some reason it's labled t2i, but it's not. it's working only in the image to image tab, not in the text to image tab. Use the Kohya one (which is stangly named anima, it's a control net why even use a style in it's name???) if you want t2i,

[@hogstoothairsoft1967]

Thanks 4 the Video... But now the third SDXL/automatic1111/controlnet video that isn't working for me, always got: "ControlNet - WARNING - ControlNet does not support SDXL -- disabling" Also tried deleting venv folder and complete reinstall... *cry* , can make it work on SD 1.5 Models, but that's not my goal... Any ideas?

[@LaCarnevali]

have you updated controlnet?

[@ronbere]

Released several days ago

[@androidgamerxc]

ok this is unrealted but i like you like i like your face and you are very beautiful not many people would find you attractive but few would go abso crazy for you and im tthatt tone

[@XDarkXThanatosX]

??? Was that a compliment or an insult. If I can't tell, then maybe she can't either. Just saying..."Not many people would find you attractive". That's fuqed up. 🤨

[@LaCarnevali]

😂😂😂

[@androidgamerxc]

@@LaCarnevali i found your Instagram you are soooooooooo beautiful i will make your fan art and tag you init

[@LaCarnevali]

please do it! @@androidgamerxc