Installation

Installation#

This document provides instructions for installing OpenFL; either in a Python virtual environment or as a docker container.

Requirements#

For a list of Python supported versions see Python version support policy. Docker images are only supported on linux/x86_64 platforms.

Using pip#

We recommend using a Python virtual environment. Refer to the venv installation guide for details.

  • From PyPI (latest stable release):

    pip install openfl

  • For development (editable build):

    git clone https://github.com/securefederatedai/openfl.git && cd openfl
pip install -e .

  • Nightly (from the tip of develop branch):

    pip install git+https://github.com/securefederatedai/openfl.git@develop

Verify installation using the fx --help command.

OpenFL - Open Federated Learning

BASH COMPLETE ACTIVATION

Run in terminal:
_FX_COMPLETE=bash_source fx > ~/.fx-autocomplete.sh
source ~/.fx-autocomplete.sh
If ~/.fx-autocomplete.sh already exists:
source ~/.fx-autocomplete.sh

CORRECT USAGE

fx [options] [command] [subcommand] [args]

GLOBAL OPTIONS

-l, --log-level TEXT  Logging verbosity level.
--no-warnings         Disable third-party warnings.
--help                Show this message and exit.

AVAILABLE COMMANDS
...

Using docker#

This method can be used to run federated learning experiments in an isolated environment. Install and verify installation of Docker engine on all nodes in the federation. Refer to the Docker installation guide for details.

  • Pull the latest image:

    docker pull ghcr.io/securefederatedai/openfl:latest

    Note

    This command works only from linux/x86_64 platforms since the image is built for this platform.

  • Build from source:

    git clone https://github.com/securefederatedai/openfl.git && cd openfl
git checkout develop

    docker build -t openfl -f openfl-docker/Dockerfile.base .

    Note

    This command copies current context (i.e. OpenFL root directory) to the base image. Ensure that the .dockerignore file is configured to exclude unnecessary files and directories (like secrets or local virtual environments).

Contents