How to Self-Host Stable Diffusion WebUI (AUTOMATIC1111) for AI Image Generation

Clive Lancaster

Clive Lancaster

— 3 min read

Stable Diffusion has revolutionized the world of AI-generated art, offering users the ability to create stunning images from text prompts. If you're interested in hosting your own instance of the popular Stable Diffusion WebUI developed by AUTOMATIC1111, this guide will walk you through the process step-by-step.

By self-hosting, you gain full control over your setup, ensuring privacy and flexibility without relying on third-party services. Let’s dive into how you can set up your own Stable Diffusion server!

Table of Contents

  1. Prerequisites
  2. Step 1: Set Up Your Environment
  3. Step 2: Install Dependencies
  4. Step 3: Download Stable Diffusion Models
  5. Step 4: Launch the WebUI
  6. Step 5: Access the WebUI
  7. Optional: Improve Performance with CUDA/GPU Support
  8. Conclusion

Prerequisites

Before we begin, ensure that you have the following:

  • A computer or server with sufficient hardware resources:
    • CPU: At least 8 cores (more is better).
    • RAM: At least 16 GB (32 GB recommended).
    • Storage: At least 20 GB of free disk space.
    • GPU (optional but highly recommended): NVIDIA GPU with at least 6 GB VRAM for accelerated performance.
  • A stable internet connection.
  • Basic familiarity with command-line interfaces.
  • Python 3.10+ installed on your system.

Step 1: Set Up Your Environment

Choose Your Operating System

You can run Stable Diffusion WebUI on Windows, macOS, or Linux. For simplicity, we'll focus on Linux (Ubuntu/Debian-based distributions) in this guide. However, similar steps apply to other operating systems.

Clone the AUTOMATIC1111 Repository:

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

Install Git (if not already installed):

sudo apt update
sudo apt install git

Step 2: Install Dependencies

The WebUI requires several dependencies to function properly. Follow these steps to install them:

Install Required Python Packages:

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
pip install -r requirements.txt
Note: The cu118 URL above assumes you're using an NVIDIA GPU with CUDA 11.8 support. If you don't have a GPU, omit the --index-url flag.

Create a Virtual Environment (optional but recommended):

python3 -m venv venv
source venv/bin/activate

Install Python and Pip:

sudo apt install python3 python3-pip

Step 3: Download Stable Diffusion Models

To generate images, you need a pre-trained model. AUTOMATIC1111 provides a script to download official models.

  1. Alternative: Manually Download Models:
    If you prefer downloading manually, visit the Hugging Face Model Hub and download the desired model files. Place them in the models/Stable-diffusion folder within the WebUI directory.

Download a Model:
Run the following command inside the stable-diffusion-webui directory:

./webui.sh

This script will prompt you to select and download a model. Choose one based on your use case (e.g., SD 1.5, SDXL).

Step 4: Launch the WebUI

Once everything is set up, it's time to launch the WebUI.

  1. Wait for Initialization:
    The first launch may take some time as the model is loaded into memory. Once ready, the script will display a local URL where you can access the WebUI.

Run the WebUI Script:

./webui.sh

This script initializes the environment, loads the model, and starts the web interface.

Step 5: Access the WebUI

After launching the WebUI, open a browser and navigate to the provided URL (usually http://localhost:7860). You should see the user-friendly interface for generating images.

Key Features of the WebUI:

  • Text-to-Image Generation: Create images from textual descriptions.
  • Image-to-Image Editing: Modify existing images.
  • Inpainting: Replace parts of an image with new content.
  • Advanced Settings: Fine-tune parameters like resolution, steps, and sampling methods.

Optional: Improve Performance with CUDA/GPU Support

If you have an NVIDIA GPU, enabling CUDA acceleration can significantly speed up generation times.

  1. Install PyTorch with CUDA Support:
    During the dependency installation step, make sure to include the --index-url flag pointing to the appropriate PyTorch CUDA repository.
  2. Test GPU Availability:
    After launching the WebUI, verify that the GPU is being used by checking the console output.

Verify CUDA Installation:
Ensure that the correct version of CUDA is installed on your system. You can check this by running:

nvidia-smi

Conclusion

Self-hosting the Stable Diffusion WebUI (AUTOMATIC1111) empowers you to explore the vast possibilities of AI-generated art while maintaining full control over your data and resources. With the steps outlined in this guide, you should now have a functional instance up and running.

Feel free to experiment with different models, settings, and extensions to enhance your experience. Happy creating!

If you encounter any issues during setup, refer to the official documentation or join the community for support.

Read more