Stable Diffusion Web-UI: A Guide to AI Image Generation

Table of Contents

What is Stable Diffusion Web-UI?
#

Stable Diffusion Web-UI (SD WebUI) is an open-source software developed by AUTOMATIC1111, based on the diffusion model originally published by CompVis in Germany. It enables image generation using AI technology through a graphical interface accessible via a web page. SD WebUI can generate images based on descriptive text inputs, mimic existing images, and even fill in content for partially blacked-out images. Its key features include:

Free and Open-Source: SD WebUI can be run for free on your own computer or server, with no usage restrictions as long as you comply with the Creative ML OpenRAIL-M licensing terms.
Easy to Use: Its graphical interface is accessed through a web page, making it user-friendly.

System Requirements
#

To run Stable Diffusion Web-UI, ensure your system meets the following requirements:

Memory: Minimum 8GB RAM, recommended 16GB.
Graphics Card: Requires a dedicated graphics card with minimum VRAM 4GB.
Disk Space: 20GB.

Stable Diffusion Web-UI primarily relies on the graphics card for computation, so any 64-bit x86 architecture Intel/AMD processor will suffice. However, powerful graphics card performance is crucial for AI image generation.

Installing Required Tools
#

Before proceeding with the installation, make sure to install Git and Python. These tools are essential for cloning the repository and building Stable Diffusion on your machine.

Use Scoop to easily set up Python:

scoop install python310 git

Install Stable Diffusion
#

Clone the repository from GitHub based on your GPU type:

For AMDGPU users:

git clone https://github.com/lshqqytiger/stable-diffusion-webui-directml.git

For NvidiaGPU users:

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

Navigate to the cloned directory and run the batch script:

cd stable-diffusion-webui
.\webui-user.bat

Wait for the installation to finish, and it will automatically redirect to the localhost offline page (your web-UI page).

Stable Diffusion Checkpoints and Lora
#

Stable Diffusion checkpoints are pre-trained models used for AI-generated image creation. They serve as the foundation for the image generation process. Additionally, Lora models, being compact yet powerful, enhance the performance and accessibility of Stable Diffusion models.

Import Models
#

Start by downloading models for image generation. Models can be downloaded from:

Place downloaded files in the following directories:

For Lora: ./stable-diffusion-webui-directml/models/Lora
For Stable Diffusion Checkpoint: stable-diffusion-webui-directml/models/Stable-diffusion

After placing the files, refresh the web-UI page to load the new models.