AI Art Generation Handbook/How to install Stable Diffusion/SD.Next

Steps for SD.Next Installation for   (Updated August 2023)

(1) Ensure your PC meet the basic requirements

(2) Download Python 3.10.10 ← Link here

(3) During the first page of installation wizard, remember to check Add Python 3.10 To PATH

(4) Proceed the rest of installation as per usual

(5) Download Git Version Control ← Link here

(6) Install using the default settings

(7) Go create a folder at your desired locations (Make sure that you have access privilege)

(8) At the address bar, type cmd

(9) At Command Shell window shall prompt and type git clone https://github.com/vladmandic/automatic

(10) Wait a moment for git to clone the whole folders

(11) Go inside the folders , and search for webui.bat

(12) Wait for a while to download the whole modules

(13) When asked Download default model (y/n) ? Press y if you do not have models (Model 1.5 not SDXL) available or n to download your preferred SD Models from huggingface/Civitai.

(14) You should see this screen below once finished.

Optional

Download improved VAE to lower down VRAM requirements : https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl_vae.safetensors

Additional Steps to Integrate SDXL into SD.Next (See screenshot for more clarifications )

(15a) In the main tab , click on Settings top tab

(15b) Click User Interface at the right side ,

(15c) Scroll down to the Quicksettings list .

(15d) Add in by typing

sd_model_checkpoint , sd_model_refiner , diffuser pipeline and sd_backend .

Note: For SDXL, recommended to go here to download the Refiner models: https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0

(15e) Scroll back to top to click on Apply Settings .

(15f) Click Restart Server button . (But highly suggest to close all of terminal windows ) and re-click webui.bat

(15g) Once clicked, you should be re-directed back to the main page .