Fooocus is an open-source, offline image generation software designed for users who want high-quality results with minimal prompt engineering. It simplifies the process, allowing users to focus on creative prompting rather than complex parameter tuning, making it suitable for both beginners and experienced artists.

How It Works

Fooocus is built on the Stable Diffusion XL architecture, enhanced with a GPT-2 based prompt processing engine and advanced sampling techniques. This approach aims to produce aesthetically pleasing images regardless of prompt complexity. It features custom algorithms for inpainting and image prompting, designed to yield more satisfying results than standard Stable Diffusion XL implementations.

Quick Start & Requirements

• Install: Download and run run.bat (Windows) or entry_with_update.py (Linux/Mac).
• Prerequisites: Nvidia GPU with 4GB VRAM and 8GB system RAM minimum. Windows users may need to enable Virtual Swap. Linux requires Python 3.10+ and Conda/venv. Mac requires macOS Catalina+ and PyTorch with MPS support.
• Setup: Automatic model downloads occur on first launch.
• Links: Install Fooocus, Colab, Linux, Mac

Highlighted Details

• Minimal GPU memory requirement of 4GB.
• Custom inpaint and image prompt algorithms for improved results.
• GPT-2 based prompt expansion for enhanced prompt understanding.
• Supports multiple presets (General, Anime, Realistic) with dedicated launchers.
• Offers advanced features like wildcards, array processing, and inline LoRAs.

Maintenance & Community

The project is currently in Limited Long-Term Support (LTS) with bug fixes only, with no current plans to migrate to newer model architectures. Community contributions are welcome, and forks are available.

Licensing & Compatibility

Fooocus is open-source and free for non-commercial use. The specific license is not explicitly stated in the README, but the project emphasizes its 100% non-commercial nature.

Limitations & Caveats

The project is in LTS with bug fixes only, meaning new features or model architecture updates are not planned. AMD GPU support is in beta and not intensively tested. Mac performance is significantly slower than dedicated Nvidia GPUs.