How to Install Stable Diffusion on Windows | 2023 Beginner's Guide | Automatic1111 Web UI

-------------------------------------------------------------
In this video, I'll guide you on how to install and use the Automatic1111 Stable Diffusion Web UI on a local Windows machine. I'll cover how to use the top Stable Diffusion models from CivitAI.com, and demonstrate how to easily recreate the example images. You'll also learn how to automatically update Stable Diffusion, its models, and extensions, as well as how to use it if you have less than 8 gigabytes of video RAM. This tutorial is designed for beginners. I've tried to make it foolproof and it's updated for July 2023. If you run into any issues during the installation, feel free to ask in the comments section.

URLs:

Timestamps:
0:00 - Introduction & Requirements
0:16 - Install Python
0:57 - Install Git
1:19 - Install A1111 Web UI Launcher
3:21 - Download & Install Additional Models from CivitAI.com
5:05 - Install Extensions and More Models
6:38 - Prompt Tweaking

[ShawnKirk]

Git page says this installer is now 'obsolete' and to use 'Stability Matrix' instead. It takes care of Python and Git for you. The only difference is the folder locations you put the sdl_base, refiner, and vae files. But it was pretty straight forward. I'm up and running now. Thanks!

[asr]

You deserve more views! Thank you for this concise guide

[Utilisateur-]

You are mentioned on a vidéo. You make a so great content ! Thanks !

[UpInSmouke]

Hi!
In the launch window, after the "Exit" button, the inscription "NVIDIA GeForce GTX 1060 3GB" is indicated. But my PC has a 4070Ti. Is this a problem or not?

[neehime]

Does this work with amd cards? If not, do you have a tuto for amd grapich cards?

[PatrickMuello]

Hello @AI Marketing Teacher I am having this error
"[notice] A new release of pip available: 22.2.1 -> 23.0.1
[notice] To update, run: -m pip install --upgrade pip

Press any key to continue . . .

Can you help me to get through this?

[livewireprod.]

I keep getting this error when I try to generate and image.

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.

I have 6gb of vram and I have selected the med vram option when loading A1111 launcher. I even checked the "cross attention to float32 option" and it didnt do anything. I havent tried the commandline option yet bc i dont know what to do.

[thaidogooder]

Can I install and run SDXL on a virtual machine? How to configure the VM for fully access to the GPU VRAM & driver?

[almas]

Is GTX 1050 Ti good enough for this to work?