Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/lllyasviel/Fooocus/llms.txt

Use this file to discover all available pages before exploring further.

Linux installations give you full control over the Python environment and are well-suited for headless servers and remote workflows. Fooocus supports three setup paths — Anaconda (recommended for most users), Python virtual environments, and native system Python — as well as AMD GPUs via ROCm. All three methods require Python 3.10.

​
Hardware Requirements

GPUMinimal GPU MemoryMinimal System MemoryNotes
Nvidia RTX 4XXX4 GB8 GBFastest
Nvidia RTX 3XXX4 GB8 GBUsually faster than RTX 2XXX
Nvidia RTX 2XXX4 GB8 GBUsually faster than GTX 1XXX
Nvidia GTX 1XXX8 GB (6 GB uncertain)8 GBOnly marginally faster than CPU
Nvidia GTX 9XX8 GB8 GBFaster or slower than CPU
AMD GPU (ROCm)8 GB8 GB~1.5× slower than Nvidia RTX 3XXX

​
Installation

Recommended if you manage multiple Python projects or plan to follow the Mac guide as well. Requires Anaconda or Miniconda.1. Clone the repository and create the environment:
git clone https://github.com/lllyasviel/Fooocus.git
cd Fooocus
conda env create -f environment.yaml
conda activate fooocus
pip install -r requirements_versions.txt
2. Launch Fooocus:
conda activate fooocus
python entry_with_update.py
On first launch, Fooocus automatically downloads the default SDXL checkpoint to Fooocus/models/checkpoints. The web UI opens at http://127.0.0.1:7865 once the download completes.

​
Model Presets

You can launch with an alternate preset using the --preset flag: 
python entry_with_update.py --preset anime

python entry_with_update.py --preset realistic
Preset flagMain modelEquivalent Windows file
(none)juggernautXL_v8Rundiffusionrun.bat
--preset animeanimaPencilXL_v500run_anime.bat
--preset realisticrealisticStockPhoto_v20run_realistic.bat

​
Remote Access

To expose the UI on your local network or a remote machine, pass the --listen flag. You can optionally specify a port with --port: 
python entry_with_update.py --listen

python entry_with_update.py --listen --port 8888
Remote access is unauthenticated by default. To add basic authentication, create an auth.json file in the Fooocus directory containing a list of {"user": "...", "pass": "..."} objects.

​
AMD GPU Support (ROCm)

AMD ROCm support on Linux is in beta and approximately 1.5× slower than an Nvidia RTX 3XXX. ROCm for Windows is currently on hold.
After installing Fooocus’s dependencies using any of the methods above, replace the default PyTorch build with the ROCm-enabled version: 
pip uninstall torch torchvision torchaudio torchtext functorch xformers
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm5.6
Then launch normally: 
python entry_with_update.py
You can still use the --preset flag alongside ROCm: 
python entry_with_update.py --preset anime

python entry_with_update.py --preset realistic

Build docs developers (and LLMs) love

Get started for freeTalk to us