So, my computer found its self-destruct button so now I need to install auto1111 on a new device. I don't have any luck finding a one click install so this is what to do if you are trying to start from scratch. A few things have changed from when I originally downloaded the program, and I thought it was annoying switching from guide to guide to find all the hidden phantom steps, so this is how I recommend downloading and some of the features I add for functionality.
The Basic installs
Before anything you need to have an Nvidia card for any of this to work.
First You need to install python and git it doesn't matter which one you install first. Git should be installed as its 64 version and python needs to be downloaded as its 3.10.6 version for compatibility. If there is a newer version of Python available kick it to the curb it doesn't work with automatic1111.
Python Release Python 3.10.6 | Python.org
Git - Downloading Package (git-scm.com)
Once python is installed you need to add it to path so you don't get errors later. For this do the following:
-open your computers "edit the system environment variables"
-click environment variables
-click path then new and edit
-put your python location
-click ok until all windows close
--don't click the x in the top right.
Next to avoid the Cuda out of memory issue download the visual studio 2022 select python and c++ when prompted and wait for that to finish.
Download Visual Studio Tools - Install Free for Windows, Mac, Linux (microsoft.com)
Once Done download Cuda for Nvidia answer the questions and it will give you the version for your device.
CUDA Toolkit 12.2 Update 2 Downloads | NVIDIA Developer
Once you have all of those, you're ready to start the actual installation process.
Create a new folder on your device, name it something you're willing to type like AI or Auto1111. Open Git Bash In the window type: cd [drive]:\[folder] ex C\:Auto1111. If your name has spaces, use "" marks around everything. Ex: "cd C:\Stable Diffusion"
On stable-diffusion-webui's GitHub, click the green "Code" button. Then click the button that copies the HTTPS (it looks like 2 overlapping squares), or copy it manually
In Git Bash, type: git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui
this should begin the main downloading process.
Now once downloaded you can use the "webui-user.bat" file to download and set up the rest of it but before that some edits should be made to it. Right click and click edit if its not in the first menu click more options then click edit if your system doesn't show either open it with notepad. When downloaded it will look like this:
Now I tend to add 3 changes, 1 is xformers which helps generation run faster, 2 is medvram so your computer doesn't use all of its resources making it more efficient but a little slower if you still get cuda out of memory errors you may need to switch to lowvram instead, and 3 is git pull for automatic updates. To add these your code should look like the one below:
set COMMANDLINE_ARGS=--xformers --medvram
To install torch go to the same folder as where your webui-user.bat file is click the bar to the left of the search bar that says the current location and type cmd in the window that comes up type: pip install torch
The download will then begin.
To update pip itself put this command in the cmd window once the process is done:
python -m pip install --upgrade pip
So, your stable diffusion should be able to boot up now just by clicking "webui-user.bat" as of writing this that means your stable diffusion should at least be version 1.6 and in my previous articles some extensions I recommended became unmaintained as of 1.5 so below is a list of workable extension resources I recommend adding to your setup. Go to the extensions tab, install from url, and put these extensions in you can read there github pages to find about more on each individually.
GitHub - DominikDoom/a1111-sd-webui-tagcomplete: Booru style tag autocompletion for AUTOMATIC1111's Stable Diffusion web UI -Auto tag completer so I can be lazy.
GitHub - pkuliyi2015/multidiffusion-upscaler-for-automatic1111: Tiled Diffusion and VAE optimize, licensed under CC BY-NC-SA 4.0 -Efficient generation resource so computer can be lazy.
GitHub - kohya-ss/sd-webui-additional-networks -Allows you to review lora info (since I always forget my own training values).
GitHub - Mikubill/sd-webui-controlnet: WebUI extension for ControlNet -For poses and a lot of other neat features (I don't use it much, but it is a popular extension).
GitHub - civitai/sd_civitai_extension: All of the Civitai models inside Automatic 1111 Stable Diffusion Web UI -The civitai extension.
GitHub - picobyte/stable-diffusion-webui-wd14-tagger: Labeling extension for Automatic1111's Web UI -An auto tagger to assist with lora making.
|Now in your settings tab click show all pages and ctrl+f and type "quicksettings" the option should look like this:
(setting entries that appear at the top of page rather than in settings tab) (requires Reload UI)
Add the "sd_vae" option from the list this makes it so you can change the active vae and checkpoint in the same way.
Now that is everything your setup should be up and running and has some nice features added as well. So, now you can go download all the lora checkpoints (ckpt) lora and vae you want to get generating below are some recommendations, don't be afraid to ignore these and use whatever you want instead.
AnyLoRA - Checkpoint - bakedVae (blessed) fp16 NOT-PRUNED | Stable Diffusion Checkpoint | Civitai -Great for lora testing
CuteYukiMix(特化可爱风格adorable style） - MidChapter | Stable Diffusion Checkpoint | Civitai -Its cute
To use these, go to your stable diffusion folder like this "[drive]:\[wherever you put it]\stable-diffusion-webui\models\Stable-diffusion" and drop in the files.
Now there are so many of these of all types of things there is no way something your interested doesn't exist. Search around and try to find ones you like. To add them though do this "[drive]:\[wherever you put it]\stable-diffusion-webui\models\lora" and drop them in.
I personally still use the orange mix vae for everything, so a link to that and the checkpoints it was originally made for is below:
WarriorMama777/OrangeMixs at main (huggingface.co)
To use a vae put the file in your folder like so, "[drive]:\[wherever you put it]\stable-diffusion-webui\models\VAE".
Okay that's everything I have on this if you have any questions feel free to ask and enjoy generating.