Stable Diffusion WebUI running either locally (CPU or GPU) or behind a Nginx reverse proxy with Let's Encrypt (ideal for installing on a headless GPU server).
It installs Stable Diffusion WebUI with xformers enabled. The default model installed is Stable Diffusion 1.5. The initial build is quite long (4-5 minutes) even on a powerful server. Subsequent launch will take a few seconds.
It's up to you to install additional models, LoRa, VAE, etc.
After cloning the repository, run:
# Initial download of models (run once)
$ docker compose --profile download up --build
# Start WebUI
$ docker compose --profile local-cpu up --build
Access from http://localhost:7860/
Ensure your system has NVIDIA Docker Toolkit installed and configured for Docker to access the GPU.
After cloning the repository, run:
# Initial download of models (run once)
$ docker compose --profile download up --build
# Start WebUI with GPU support
$ docker compose --profile local-gpu up --build
Access from http://localhost:7860/
After cloning the repository, create .env file and customize with your own domain and email:
VIRTUAL_HOST=sd.yourdomain.com
VIRTUAL_PORT=7860
LETSENCRYPT_HOST=sd.yourdomain.com
[email protected]
Before proceeding, ensure that your DNS settings are correctly configured to point your domain (e.g., sd.yourdomain.com) to the IP address of your cloud server. This step is crucial for Let's Encrypt to verify your domain and issue a SSL certificate.
Create .htpasswd file to setup HTTP basic auth to protect the Web UI:
$ htpasswd -c .htpasswd username
You may need to install apache2-utils package:
$ sudo apt-get install apache2-utils
You are ready to build & launch:
# Initial download of models (run once)
$ docker compose --profile download up --build
# Start WebUI in detached mode (-d)
$ docker compose --profile cloud up -d --build
The initial build is quite long (4-5 minutes) even on a powerful server. Subsequent launch will take a few seconds.
The WebUI will be available at https://sd.yourdomain.com (as defined in your .env file).
If you have any issue, you can check the logs:
$ docker compose logs -tf
All persistent data including models, extensions, and settings are stored in a Docker volume named data. This ensures your configurations and installed components persist across container restarts and updates.
Generated content (images, etc.) is stored in a local outputs directory, which is mapped to the container. This makes it easy to access and manage your generated content from your host system.
Compatibility matrix for PyTorch
https://github.com/pytorch/pytorch/wiki/PyTorch-Versions
Thanks to:
- Stability.ai for the fairly open image diffusion model
- AUTOMATIC1111 for the WebUI
- AbdBarho for the download container idea
- and the huge SD community for all the tutorials and models!