Fooocus

Overview

Fooocus is an image generation software that brings Midjourney-like simplicity to Stable Diffusion XL. Created by the developer of ControlNet, it streamlines the complex parameters of image generation so users can focus on prompts and images. Fooocus includes built-in prompt expansion using GPT-2, various artistic styles, and advanced features like image prompts and inpainting while maintaining a user-friendly interface.

System Requirements

• CPU: Modern multi-core processor (Intel/AMD)
• RAM: 16GB+ recommended
• GPU: NVIDIA GPU with 4GB+ VRAM (8GB+ recommended for optimal performance)
• Storage: 15GB+ for application and models
• OS: Windows, Linux, or macOS (including Apple Silicon via MPS)

Installation Guide

Option 1: Windows Installation (Easiest)

1. Download the latest Windows package from the GitHub releases page
2. Extract the ZIP file using 7-Zip (right-click file → properties → unblock if needed)
3. Run one of the launcher batch files:
   • run.bat - Standard edition
   • run_anime.bat - Anime-optimized edition
   • run_realistic.bat - Realistic-optimized edition
4. The interface will open in your default web browser automatically (usually at http://127.0.0.1:7865)

Option 2: Manual Installation (Windows/Linux/Mac)

For Windows

1. Clone the repository:

   git clone https://github.com/lllyasviel/Fooocus.git
   

2. Navigate to the project directory:

   cd Fooocus
   

3. Run the update script to download models and dependencies:

   python entry_with_update.py
   

For Linux

1. Ensure Python 3.10 is installed

2. Clone the repository:

   git clone https://github.com/lllyasviel/Fooocus.git
   cd Fooocus
   

3. Set up a Python virtual environment:

   python3 -m venv fooocus_env
   source fooocus_env/bin/activate
   pip install -r requirements_versions.txt
   

4. Run the application:

   python entry_with_update.py
   

For macOS (including Apple Silicon)

1. Install Conda and PyTorch nightly

2. Clone the repository:

   git clone https://github.com/lllyasviel/Fooocus.git
   cd Fooocus
   

3. Create and configure the Conda environment:

   conda env create -f environment.yaml
   conda activate fooocus
   pip install pygit2==1.12.2
   

4. Run the application:

   python entry_with_update.py
   

Option 3: Google Colab (No Installation)

If you don’t have a suitable GPU, you can run Fooocus on Google Colab:

1. Visit colab.research.google.com

2. Open a new notebook

3. Run the following commands:

   !pip install -q https://github.com/camenduru/stable-diffusion-webui-colab/releases/download/0.0.20/xformers-0.0.20+cu118.pt.whl
   !git clone https://github.com/lllyasviel/Fooocus
   %cd Fooocus
   !pip install -r requirements_versions.txt
   !python entry_with_update.py --share --always-high-vram
   

4. Click the link generated with “gradio.app” to access the interface

Note: The first run will download required models (approximately 7GB).

Practical Exercise: Getting Started with Fooocus

Let’s walk through a simple exercise to help you get familiar with Fooocus’s capabilities.

Step 1: Basic Text-to-Image Generation

1. Launch Fooocus and wait for the interface to load in your browser
2. In the main text field, enter a simple prompt like: “forest elf, fantasy character, detailed, colorful clothing”
3. Click “Generate” to create your first image
4. Notice how Fooocus automatically enhances your prompt in the background (you may see the expanded prompt in the console)
5. The generated image will appear in the gallery below

Step 2: Using Styles and Presets

1. Scroll down to the “Style” section
2. Try selecting different style options like “Cinematic,” “Photographic,” or “Fooocus V2” (the default)
3. You can select multiple styles simultaneously - try combining “Fooocus V2” with “Cinematic”
4. Generate a new image with the same prompt to see how styles affect the output
5. Try switching to a different preset using the dropdown menu at the top (Default/Anime/Realistic)

Step 3: Advanced Features

Using Image Prompts

1. Click the “Image Prompt” tab
2. Upload a reference image by clicking “Upload”
3. Enter a text prompt that describes additional elements or styles
4. Adjust the weight slider to control how much influence the image has
5. Click “Generate” to create an image based on both text and image prompts

Inpainting

1. Click the “Inpaint” tab
2. Upload an image you want to modify
3. Use the brush tool to paint over areas you want to change
4. Enter a text prompt describing what should appear in the masked area
5. Click “Generate” to create the inpainted result

Step 4: Fine-Tuning Parameters

1. Click “Advanced” to expand the advanced settings

2. Experiment with these parameters:

   • Sampling Sharpness: Affects texture detail (try 2.0 - 7.0)
   • Guidance Scale: Controls adherence to prompt (try 4.0 - 9.0)
   • Quality/Speed: Adjusts generation steps (higher is better quality but slower)
   • Seed: Set a specific seed for reproducible results

3. Generate images with different parameter combinations to understand their effects

Step 5: Saving and Sharing

1. When you create an image you like, right-click on it and select “Save Image”
2. The saved PNG file contains embedded metadata with generation parameters
3. You can drag and drop this PNG back into Fooocus later to load all the same settings

Resources

Official Documentation

• GitHub Repository README - Main documentation and installation guides
• Advanced Features Guide - Detailed explanation of all features
• Community Generated Examples - Non-cherry-picked examples showing Fooocus capabilities

Community Support

• GitHub Discussions - Community support and showcase
• GitHub Issues - Bug reports and feature requests
• Matrix Space: #comfyui_space:matrix.org - Related community chat for broader Stable Diffusion discussion

Related Projects

• Stable Diffusion XL - The base model used by Fooocus
• ControlNet - Another project by the same developer
• ComfyUI - Node-based interface that shares some code with Fooocus

Tutorials and Guides

• Image Prompt Tutorial - Guide for using image prompts effectively
• Style Comparison Guide - Visual examples of different style settings
• LCM & Advanced Settings - Explanation of LCM and other advanced settings