Installation

This page describes how to install APPFL on a machine independent to operating systems. Machine-specific installation steps may be added later.

Conda environment

We highly recommend to create new Conda virtual environment and install the required packages for APPFL.

$ conda create -n APPFL python=3.8
$ conda activate APPFL

User installation

For most users, including data scientists, this simple installation must be sufficient for running the package.

$ pip install pip --upgrade
$ pip install "appfl[analytics]"

If you want to even minimize the installation of package dependencies, you can use:

$ pip install appfl

Note

torch may need to be updated manually to supprt CUDA. Please check GPU support in PyTorch.

Developer installation

Code developers and contributors may want to work on the local repository. To set up the development environment,

$ git clone https://github.com/APPFL/APPFL.git
$ cd APPFL
$ pip install -e ".[dev,examples,analytics]"