Updated Stable Diffusion A1111 2023 Installation Tutorial V1.5, V2.1, and Custom Models
Vložit
- čas přidán 2. 06. 2024
- SDXL is out and the only thing you will do differently is put the SDXL Base mode v1.0 and Refiner Model v1.0 into your model's folder the same as you would with any model (i.e. that's where your 1.5 and 2.1 model went) and then generate txt2img using the base model. To use the refiner send your generated image to img2img, choose the refiner model, match the resolution to the txt2img pic, and then set a low denoise (0.1-0.3) and generate.
Here are thew new SDXL models.
1. sd_xl_base_1.0_safetensors - huggingface.co/stabilityai/st...
2. sd_xl_refinder_1.0_safetensors - huggingface.co/stabilityai/st...
This is an updated guide for Installing Stable Diffusion Automatic 1111 (A1111) v1.5, v2.1, and Custom models. Some things have changed recently and installing will slightly easier now we don't have to add yaml config files manually.
Things you will install in order:
Step 1. Install Python 3.10.6 & GIT (do not download updated versions of python)
Python 3.10.6 - www.python.org/downloads/rele...
GIT - git-scm.com/downloads
Step 2. Download A1111 repository & models
github.com/AUTOMATIC1111/stab...
V1.5 - huggingface.co/runwayml/stabl...
V2.1 - huggingface.co/stabilityai/st...
Great Site for Custom Models -civitai.com/
Optional Install - Cuda Drivers 11.8 (compatible with pytorch) - developer.nvidia.com/cuda-11-...
*IMPORTANT if you update your stable diffusion by overwriting the newest repository over your current Stable Diffusion Directory, save a copy of your webui-user.bat file or make sure you restore the --xformers or other ARGS that you had before the overwrite.
Do you have a MAC? Check out this awesome tutorial stable-diffusion-art.com/inst...
Do you have a MAC? Want the standard tutorial? Here it is github.com/AUTOMATIC1111/stab...
Do you have an AMD? Try this tutorial -github.com/AUTOMATIC1111/stab...
Did Pytorch stop you? Install is manually by getting the command to download it through python - pytorch.org/
Choose pip for package & 11.8 for CUDA
Some useful ARGS
--xformers
--medvram
--no-half
--precision full
List of ARGS and definitions (Arguments i.e. --xformers) - github.com/AUTOMATIC1111/stab...
Awesome Stable Diffusion site w/ guides for everything (not my site) - stable-diffusion-art.com/begi...
Nvidia GPU Architecture List - en.wikipedia.org/wiki/List_of...
More info about Xformers - github.com/AUTOMATIC1111/stab...
==========================
Chapters
00:00 Intro: What has changed & what will be in this video
00:36 PreReqs
00:55 Install Python 3.10.6, GIT, & optional CUDA
02:13 Download Automatic 1111 GitHub Repository to your PC
03:30 Download V1.5, 2.1, & Custom Model & What's EMA vs. Non EMA and Safetensor vs. CKPT
05:40 Bring your Models into Stable diffusion/Models/Stable-Diffusion Folder
06:25 Run Webui-user.bat file to Start installing Dependencies 10-15min
07:43 Editing your Webui-user.bat File to Add ARGS to fix V2.1 and Add --xformers
09:23 How do you run this thing and generate images?
09:50 Txt2Img Tab the Main tab to create images overview
11:38 Example of Restore Faces + Hires Fix
12:04 CFG Scale
12:28 Seeds
12:49 Img2img
13:05 Denoising
13:43 Inpainting
14:12 What is CLIP?
14:42 Extras Tab (Upscaling and GFPGAN) & Example
Seems like I didn't go over upscaling in the extras tab. It increases the quality and resolution of your base image, just throw it in there and select the number of times you want to increase the resolution. I personally do a latent upscale instead which uses Img2Img w/ denoising under 0.40 and I double the resolution. It will recreate some of the finer details and make the picture look even better.
#Automatic1111 #A1111 #StableDiffusion #StableDiffusiontutorial #stablediffusionaitutorial #aiart #aiartcommunity - Věda a technologie
![Stable Diffusion Tutorial - How to Install, How to Use [Nvidia & AMD]](http://i.ytimg.com/vi/i0_QKju-3PQ/mqdefault.jpg)
The production on your videos is just top-notch. Great video making and great tutorials.
Fantastic video and content mate,keep it up ! Subbed !
Thanks for the sub! And glad u liked the vid.
Great tutorial! Do you happen to know if Stable Diffusion is private, or if they take our photos (generated or not) to do what they wish with it?
I can't believe i finally found a tutorial that worked. awesome and thank you.
Awesome! Welcome to stable diffusion and automatic 1111 community then.
thx for this!🙂
No prob!
Damn! So good.
Thanks for the kind words!
thanks for the video my man, new sub! could you make a video for Dreamboth? i have some issues with model training (Exception training model: 'No executable batch size found, reached zero.')
Hi and thanks! Definitely doing dreambooth at some point but they haven't seemed to optimize it and LoRA can do what dreambooth does pretty good. I was reading this blog and some guy fixed that error but I didn't understand the fix, github.com/d8ahazard/sd_dreambooth_extension/discussions/787
Hey thanks for this! So i was having problem with my previous version of A111...getting error that said..
CUDA out of memory. Tried to allocate 90.00 MiB (GPU 0; 4.00 GiB total capacity; 3.29 GiB already allocated; 0 bytes free; 3.40 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF'. Check your schedules/ init values please. Also make sure you don't have a backwards slash in any of your PATHs - use / instead of . Full error message is in your terminal/ cli....
So i was frustrated looking at Discord. And then found your Video, and so much thankful to you for putting this up Fresh and Neat!
.
Just one question, i have a NVIDIA RTX 3050 with 4gb Vram. will downloading the CUDA driver you said at 1:58 help me ??
I would try it out. You can test it by upping your resolution on a generation till you get a out of vram error then try it again with the xformers and cuda drivers. My bet is it will give you a little bit more room to play with but that's a just a guess.
My understanding of the drivers is it allows utilize your cuda cores to run processes parallel to make it more efficient but it still uses VRAM. What Nvidia means by efficient and optimized wasn't clear to me. They might just mean faster. My understanding is pretty basic on CUDA drivers. I could not find info on if it uses less VRAM but the generation is definitely optimized in terms of speed so always go with CUDA and xformers if you could BUT as I was looking for more info and ran into an article that PyTorch 2.0 has an SDP (--opt-sdp-attention) but it gets confusing because A111 installs a lower version (this was the article github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/8691) . I'll have to read up on SDP or semidefinite programs. Long story short I would try and it add --xformers, but it seems there is another option now with the SDP command w/ torch/gumpy 2.0 that is just as fast.
@@ArtificallyIntelligentAi Thanks for writing back! It so happend that i was able to run A1111 after following ur tutorial, and was running on CUDA 11.8 , the one you shared. But today when i saw there's an 12.1 update, and when i updated that, it's all fucked up again. resorting back to 11.8 now.
@@cashyappurushottam427 ya the new cuda drivers aren’t optimized for PyTorch which is one of the dependencies for a1111, and perhaps the most important one. You can force it to work by installing it directly but it’s just going to give you more problems. 11.7 or 11.8 are the way to go for cuda drivers for a1111.
does this have to be on your c drive? trying to install to an external ssd
Hi, I did follow instructions and dowloaded all these files and started using automatic 1111 (my main reason was really to be able to use deforum) so I downloaded deforum into my automatic 1111 also - but not sure why I cant upload images or videos directly from my pc .. I tried clicking on the folder thumbnail but it does not work, I can get image links from online sources and paste into the image section and it works but cant upload my own images - any advice please ? thanks in advance :)
Hey there! I would make sure you click on the extensions tab and then update it via the installed sub tab> click on check for updates > apply and restart. Also most extensions won't allow you to drag stuff in if your check point models are not compatible. This happen to me when I was using a pix2pix model. Try the regular v1.5 one first. It should work with the custom models tho since they are all v1.5 base models. Try that first and see if it works. If it doesn't work, "try" to generate a video in deform without uploading anything then look at your webui-user.bat file and tell me what it says. It should give you an error that might help us figure it out. You can also scroll up to see if there are any errors at all when you started your webui-user.bat file. You can also join their discord and ask the devs and community questions directly discord.gg/deforum
You can also try removing it and installing via the install from URL tab in extensions using this link github.com/deforum-art/deforum-stable-diffusion.git , these new tools are all in beta so sometimes you just need to try a diff way of installing. A1111 might of had a bug grabbing it from extensions. It's just an easy check you can do, I'll keep an eye out for more solutions though.
@@ArtificallyIntelligentAi Ok thanks ... I re try ;)
Thanks for the video. When i do edit for installing xformer i don't see the same thing. You have like 4 line i have a loot of them. I think i got something wrong
Weird it should be the same if you got it from the A1111 link in my description. That should be the default. Some people manually put git pull to automatically updateA111 and some people have to add the python path manually because they didn't add it to path when install or add an ARG for the API. What does some of the extra stuff say? I'm sure myself or someone else that ran into that issue might be able to tell you what it is.
@@ArtificallyIntelligentAi look like the line disappeared, it working. Thanks for answering
I did all of this and mine gets to: Model loaded in 6.8s (load weights from disk: 2.2s, find config: 1.1s, create model: 0.2s, apply weights to model: 0.6s, apply half(): 0.7s, move model to device: 0.9s, load textual inversion embeddings: 1.0s). And then it just does nothing afterwards. No URL or anything. Big sadge 😢
I uninstalled the version 3.10.6 of Python and installed the latest stable release with the set to PATH checkbox checked during installation and also downloaded v1-5-pruned.ckpt from hugging face and the yaml file and moved those into my models/stablediffusion folder and then tried running it and now I get a URL. Before I was using python 3.10.6 and the 768-v-ema.ckpt file as my only checkpoint file in my models/stablediffusion folder. The terminal or whatever looks like this now: venv "C:\Users\44774\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Version: v1.2.1
Commit hash: 89f9faa63388756314e8a1d96cf86bf5e0663045
Installing requirements
Launching Web UI with arguments:
No module 'xformers'. Proceeding without it.
Loading weights [bfcaf07557] from C:\Users\44774\stable-diffusion-webui\models\Stable-diffusion\768-v-ema.ckpt
Running on local URL: 127.0.0.1:7860
To create a public link, set `share=True` in `launch()`.
Startup time: 8.1s (import torch: 1.8s, import gradio: 1.3s, import ldm: 1.1s, other imports: 1.4s, load scripts: 1.4s, create ui: 0.9s, gradio launch: 0.1s).
Creating model from config: C:\Users\44774\stable-diffusion-webui
epositories\stable-diffusion-stability-ai\configs\stable-diffusion\v2-inference-v.yaml
LatentDiffusion: Running in v-prediction mode
DiffusionWrapper has 865.91 M params.
Applying cross attention optimization (Doggettx).
Textual inversion embeddings loaded(0):
Model loaded in 8.0s (load weights from disk: 2.3s, find config: 1.2s, create model: 0.2s, apply weights to model: 0.7s, apply half(): 0.9s, move model to device: 0.9s, load textual inversion embeddings: 1.9s).
It's weird how my local URL displays in a different section than all of the tutorials I've seen, I almost missed it and thought it wasn't working still.
Would you use this update installation, to install on the google remote desktop?
You just need the new sdxl and refiner models and 1.5 is automatically downloaded for you but the rest is the same I’m making an updated tutorial soon. There is also fooocus now which is a one click installer no need to do prerequisites or download models it does it all
Thanks for the quick response, I'll look forward to updated tutorial@@ArtificallyIntelligentAi
hey there, i dont see the link for the installation on mac, i believe it is not in description tho
Sorry they are in the GitHub page but let me put it in the description now thanks for letting me know
Here is the tutorial for Mac stable-diffusion-art.com/install-windows/ this will be in the video description as well now. There is also the A1111 version of the tutorial but its less user friendly github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Installation-on-Apple-Silicon
@@ArtificallyIntelligentAi thank you ! By the way is it necessary to install stable on your mac if you want to use stable diffusion with deforum , or u can also use it through google collab ? Im just curious as i am running out of memory on mac , already installing stable out there , but maybe it is easier to run in online somehow and cleaning up the space on mac . Thanks un advance !
Hello ! I have a question with installing the webui-bat…
So, it told me to upgrade pip to its latest version which i recently just did but it continues to tell me that it’s still on version 22.2.1 ? It upgrade to 23.1 and said “successful”
After that, i went back to webui-batch and while installing, it still within seconds told me to upgrade…do you know how to fix this..?
(after trial and error): I think i was in the wrong folder…i was in another stable diffusion file ? Idk…but it seems to he installing quite better AND faster 😅 sorry
well…after installing- it gave me the error again. At the very last min of it finishing its download, it still gave me the error…could you tell me what im doing wrong? ive tried upgrading it and it says its on the latest version but still nothing…
Hi there I think I got the solution to upgrade Pip for A1111 so it doesn't just update the windows one. Go to your stable diffusion > VENV > Scripts and then right click on your Python.exe and select copy file path (win 11 show more options first) then press win Key + R to open up your run panel and type CMD and press enter to open up your command prompt then right click in your command prompt or press ctrl + V to paste that address you copied. Directly after that line press space bar and then type -m pip install --upgrade pip and press enter. This will install it directly to the python in your VENV (virtual environment). The full code for me looks like this "E:\stable-diffusion-webui\venv\Scripts\python.exe" -m pip install --upgrade pip yours will be a bit different depending on where you installed A1111. Hope this helps. Pip never stopped me from installing everything though. Let me know if you get a diff error now. Sometimes if you dont have the windows runtime C++ binaries it will throw up a bunch of "can't download" errors. So if it starts to tell you that you can't update or download something else its prob the binaries.
@@ArtificallyIntelligentAi I thank you a ton ! I was able to get A1111 up and running but - i apparently domt have enough GPU…so, i looked more into it and its bc i have an AMD…so..now time to save up for an NVIDIA 😅 again, thank you ! Its up now…i just need to change the AMD to NVIDIA..unless you know how to fix this issue? If not its okay! Thank you
@@InADifferentTime If its just low VRAM you can try vlad diffusion which is similar to A1111 i dont know if they do AMD though. You can also render with under 256*256 resolution or lower until it works and under 20 steps with Eular A. Then send it back into the extras tab in stable diffusion to up the quality again. It's going to take some testing on how high you can go till it says out of Vram, just keep a eye on your webui-user.bat cmd window if it has a progress bar, if no... go lower. You can also add --lowvram or --medvram right before your --xformers or --no-half . There are free sites though that offer almost all the same features of stable diffusion 1.5 and 2.1. Check out my Leonardo AI and czcams.com/video/QlTjGhRGJtc/video.html You can actually do training, Custom models, AI canvas, out painting, in painting. It's a pretty good amount of free stuff daily. 150 credits for leonardo ai and 1000 for playground ai and it's 1-4 credits per generation.
@@ArtificallyIntelligentAi Its odd because it loads the image tio 100% and then tells me i dont have enough GPU...ill try out your prompts in the webui-userbat. Again, sorry for the bother but thank you so much for helping...
SDXL is out and the only thing you will do differently is put the SDXL Base model v1.0 and Refiner Model v1.0 into your model's folder the same as you would with any model (i.e. that's where your 1.5 and 2.1 model went) and then generate txt2img using the base model with at least 1024 resolution. To use the refiner send your generated image to img2img, choose the refiner model, match the resolution to the txt2img pic, and then set a low denoise (0.1-0.3) and generate.
Here are thew new SDXL models.
1. sd_xl_base_1.0_safetensors - huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/tree/main
2. sd_xl_refinder_1.0_safetensors - huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/tree/main
By the way SDXL doesn't play nice with some extensions and doesn't work with LoRAs or embeddings right now as those were trained with previous models.
Stable Diffusion no longer requires Yaml files to be installed manually. Here is an updated tutorial for 2023 on how to install Stable Diffusion v1.5, v2.1, and custom models from CIVITAI and also how to edit your Webui-user.bat file with Diff ARGS to utilize xformers or stop v2.1 from running out of VRAM and rendering black/blank images.😇🍒🌶
Do you have a MAC? Check out this awesome tutorial stable-diffusion-art.com/install-mac/
Do you have a MAC? Want the standard tutorial? Here it is github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Installation-on-Apple-Silicon
Do you have an AMD? Try this tutorial -github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Install-and-Run-on-AMD-GPUs
Hi
@@LouisGedo O Hai. 00:22
Stable diffusion model failed to load
Applying optimization: sdp-no-mem... done.
where to find the other models u have there like in paint?
Here is the runwayML inpainting model huggingface.co/runwayml/stable-diffusion-inpainting/tree/main the JinxLoL one I left out because its nsfw
Hi, when I try to generate an image, I get this error: NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.
Problem solved, I have changed the model and it works as intended.
@@allmight9125 Nice.
When i run the git clone and copy paste it says, unable to access howndo i fix it
Hi there, I would just use the first method at 02:25 which is unzipping the repository. You don't have the use the git clone especially if your having issues. Permissions issues are hard to troubleshoot, they are almost like invisible errors since it doesn't throw up an error code and everyones permissions set up can vary. You can also try reinstalling or modifying python and ensuring you click on "install launcher for all users" and give it as much permissions as you can. If your on a guest account then just do the winzip method.
Mine always stops at downloading pytorch which is a 2.6 GB and stucks there, what should I do
Check if u have a c++ error at beginning of your install. You install it manually via command prompt using a code u have from this site pytorch.org/get-started/locally/
Bro webui-user-bot file is missing in github link i try tol of time but i facing same problem pls give the solution
The first thing I would do is a Ctrl+F and try and find the file that way as well, Keep in mind there is another file with the same exact name but its a shell script (.sh) not a Batch file (.bat). You can make your own .bat file by opening up a notepad and putting the webui-user.bat file info from the official site github.com/AUTOMATIC1111/stable-diffusion-webui/blob/master/webui-user.bat copy and paste that info into the notepad then save as webui-user.bat inside your Stable diffusion main folder, make sure the file extension is a .bat and not .txt. The file icon should have gears on it if you did it successfully. Also, did you add python to path and did you install version 3.10.6 and not another version?
Something you can try is installing python again but this time choose modify (since its already installed) and then checkmark add python to environment variables, and checkmark install/allow all users, you can also checkmark disable the path length limit, that's more for creators than users and it will eliminate one possible issue.
Also if you grabbed the repository using the ZIP technique try the other installing the repo the second way by directly going through python directly by calling the github repo via git clone see 02:42.Hopefully one of those suggestions work.
Thnks buddy
buddy when i running webui-user.bat file i geting Error code: 1 what is the solution
Do you have a venv (virtual environment) folder in stable diffusion? Also, does it say exit code or error code 1? And does it say access denied? If it does something is blocking it with permissions. Are you the admin on the pc? Also another possibility is a conflict between multiple versions of python. There was quite a few possibilities. The a1111 issues tab on the main GitHub has solutions and screenshots so you can compare their issue with yours. I personally just reinstall and ensure you use python 3.10.6, add python to path and use cuda 11.8 m. Basically use my links to older versions because the new ones aren’t supported. Sorry I got to deal with all this stuff. What graphics card do you have and did you know u can use stable diffusion for free on Leonardo ai or playground ai, I made a video on that.
If it says something about the file being in use, I restarted my computer and it fixed this. For some reason something was running in the background.
I had a lot trouble, I fixed all of them except this one "ImportError: DLL load failed while importing cv2: The specified module could not be found."
Please help me if someone knows how to fix this. I spends hours looking for answers but none of them work.
Sorry your having this issue and forgive me if you already tried this but Did you add python to Path when installing? It sounds like it can't find the python.dll file but that file shouldn't be missing so I'm guessing it's looking in the wrong place. You can open up your command prompt and type echo %path% and then it should show you a bunch of addresses. Look for one that says python310/scripts. You also have to be running python 3.10.6 and only that version. You can also press the windows key and type in environment variables and there should be a path there that leads to that python310/scripts, which is another way to check. If there isn't, you have to reinstall python 3.10.6 and ensure add to path is clicked. You can also try upgrading your C++ libraries learn.microsoft.com/en-us/cpp/windows/latest-supported-vc-redist?view=msvc-170 you'll need the x64 probably. These are 2 things that will prevent a lot of issues and might fix yours. You can also scroll up and see if there was another error that is more specific when your webui first started running. I had the C++ libraries one time but it was way at the top after 50 lines of code then I got another error at the bottom.
@@ArtificallyIntelligentAi ill try this thanks. Im actually using 3.11. I used 3.10.6 b4 but its asking to update something, tried to update but it wont go through so i changed it to 3.11. Lots error, i just cant fix this cv2.
ModuleNotFoundError: No module named 'font_roboto' error please fix.
Some people are saying install with the winzip method instead of git clone but seems like some of your A1111 files didn't make it over or it's corrupted. Try and install using the winzip method and replace the existing files. If you did use the winzip method just try with a fresh install and move cut and paste your models (the big 5gb files from hugging face or civit AI) into your freshly installed location. Make sure your stable diffusion isn't running when you do it.
i just download SD but a dont have the lora model file do you know why '
By default the newer automatic 1111 stable diffusion should have a Lora extension but it’s accessed via that pink icon under the generate button which is the hyper networks. Watch this video to learn about Lora and embeddings and how to get some Lora models czcams.com/video/pmdSxDSsbUs/video.html
stuck on model loaded, isn't going past that on the batch Model loaded in 13.1s (calculate hash: 8.8s, load weights from disk: 0.2s,
Im having same problem, is there a fix for this? I tried re-downloading the A1111 repository, updating PIP, and redownloading python 3.10.6
@@MeatMake Mate i swear i've been reinstalling, trying different methods all day. it's still doing the same thing :/
@@harryhealy611 i think i found a solution, if you manually type in the URL that's shown on the video into the web browser it works (7:35)
@@MeatMake i hope so! i'll try it now.
@@harryhealy611 Let me know if typing it in manually works. I'm interested to know.
Well now i know why it didnt work yesterday lol, i installed a1111 and at some point it said dowloading something but nothing happened, guess i need to install puthon and the git something lol
O ya you def need python hand make sure it’s the 3.10.6 version or some stuff won’t be compatible u also have to add python to path when installing (there will be a checkbox for it) GIT also helps u grab everything from the git hub repositories automatically. So u got it working now?
i got cuda error what should i do for that?
is it an out of memory cuda error? There is a few cuda errors. If it's that you can lower the resolution on the images you crate or type --medvram in the webui-user.bat file where you put t--xformers or --no-half. What video card do you have and what version of cuda drivers do you have. It can only take the older cuda drivers that I put in my video descriptions anything above 11.8 might give you errors. Let me know what the specific error was though and just paste that in here. Check out this video in the meantime until you get your Stable Diffusion to work czcams.com/video/QlTjGhRGJtc/video.html you can pretty much everything you can do with stable diffusion for free.
6:32
10 minutes? 10 MINUTS???
It took me 4 hours to install the damn thing! Error after error, I was forced to even change Windows. Good thing I didn't punch the monitor!
>.< that sucks, so sorry that was your experience. That happen to me when I first started because my C++ libraries weren't updated the requirements document wouldn't download automatically so it asked me to use git install for each error. But at least it's up and running now.
im getting a could not launch python error
latest version is already installed
@@anodicdubz2114 hi you have a version higher than 3.10 so it won’t install the version I linked. Newer versions are not compatible with automatic 1111 just uninstall the newer version of python completely and use the version 3.10.6 in my video description. There is an exe installer at the bottom of the python website. I recommend starting with a fresh automatic 11111 install and running the webui-user.bat file again to ensure it installs the dependencies and the venv folders correctly.
@@ArtificallyIntelligentAi actually i am on that version.
also just realized im on amd. and was following the nvidia guide
At the second run It keeps sayis installing reuirements and doesn't do anything the batch.
i.gyazo.com/32da8c8d8c3c6b0e6687701fa5c7a319.png
What kind of GPU do you have? It says torch is not compatible with your GPU. It can be a few things but it seems like PyTorch won't run. If you scroll up and python won't install then you need the C++ libraries learn.microsoft.com/en-us/cpp/windows/latest-supported-vc-redist?view=msvc-170. You, can also try installing PyTorch Directly through this site pytorch.org/ and choose the options that describe your PC and it will give you a code and just throw that code in Command Prompt and press enter and it should manually install PyTorch which "might work". It can be a number of problems. You can also throw --precision full or --no-half in your ARGS inside your webui-user.bat file, First thing I need to know is what card you have.
@@ArtificallyIntelligentAi I had 5 versions of pythong I uninstalled all and reinstalled the one of 3.10 , the one I was using was 3.11 , but anyway I did manage to load and start but , I am not really happy with the results I gues... or may be I am not able to use it properly yet, what I would like to do is to load a art style I did, and the SD dupplicate it and make variations , but so far no luck , it just creates bad looking stuff... :( I tried also to load a mask to generate art just in an area and no luck there either :/ ..