Docker Stable Diffusion WebUI Forge

Dockerized Stable Diffusion WebUI Forge

This repository provides a comprehensive Docker setup for running Stable Diffusion with a customized WebUI interface. The configuration has been optimized to support GPU acceleration via NVIDIA CUDA, ensuring high performance for generating AI-based images.

Prerequisites

• Docker and Docker Compose installed on your system.
• NVIDIA GPU with driver support for CUDA (Version 12.7 recommended).
• NVIDIA Container Toolkit for GPU support within Docker.

Project Overview

The repository is structured to build and deploy a Docker container that includes:

• CUDA 12.6.2 and cuDNN 8 for GPU acceleration.
• Python 3.10 as the runtime environment.
• PyTorch and other dependencies configured for stable operation.
• Options for persistent data storage, ensuring models and generated content are maintained across container restarts.
• Customizable launch arguments managed through the webui-user.sh script for flexibility.

Setup and Installation

1. Clone the Repository:

   git clone https://github.com/alex-des-santos/docker-stable-diffusion-webui-forge.git
   cd docker-stable-diffusion-webui-forge

2. Docker Compose Configuration: The docker-compose.yml file configures the build and runtime properties for the container:

   services:
     stable-diffusion-webui:
       image: docker-forge-webui
       build:
         context: .
         dockerfile: Dockerfile
       ports:
         - "7860:7860"
       volumes:
         - ./data:/app/data
         - ./models:/app/models
       environment:
         - DEBIAN_FRONTEND=noninteractive
         - TZ=America/Sao_Paulo
       restart: always
       deploy:
         resources:
           reservations:
             devices:
               - driver: nvidia
                 count: 1
                 capabilities: [gpu]

3. Building the Docker Image: Build the Docker image using the provided Dockerfile:

   docker-compose build

4. Running the Container: Start the container with Docker Compose:

   docker-compose up -d

Customization

Command-line Arguments

Custom arguments for the WebUI can be set in the webui-user.sh script:

#!/bin/bash
export COMMANDLINE_ARGS="--pin-shared-memory --xformers --cuda-malloc --listen"

Troubleshooting

• GPU Not Detected: Ensure that the NVIDIA Container Toolkit is correctly installed, and the Docker Desktop settings allow for GPU usage.
• Web Interface Not Accessible: Confirm that the container is listening on 0.0.0.0 and check the mapped port (default 7860).
• Permission Issues: Run the container as a non-root user if needed and set appropriate file permissions within the Docker setup.

Useful Commands

• Check Container Logs:

  docker logs <container_id>

• Access Container Shell:

  docker exec -it <container_id> /bin/bash

License

This project uses content from NVIDIA and other related packages that are governed by their respective licenses.

Contributions

Feel free to submit pull requests or raise issues to improve this repository.

Author

Developed by Alex dos Santos.

---