Skip to main content

Installation

Quickstart​

Installing using pip​

Please make sure the installed Python version in your system is one of the following: 3.8, 3.9, 3.10.

pip install fedml

Installing using conda​

In the example below, we use Python version 3.8 but feel free to use another version from the supported versions: 3.9, 3.10.

conda create --name fedml-pip python=3.8
conda activate fedml-pip
conda install --name fedml-pip pip
pip install fedml

Test installation​

If the installation is successful, you will not see any issue when importing fedml

Python 3.9.6 (default, Aug 11 2023, 19:44:49) 
[Clang 15.0.0 (clang-1500.0.40.1)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import fedml
>>>

Advcaned​

Installing additional ML engines​

The default machine learning engine is PyTorch. FedML also supports TensorFlow, Jax, and MXNet. Engines can be installed individually:

pip install "fedml[tensorflow]"
pip install "fedml[jax]"
pip install "fedml[mxnet]"

or altogether:

pip install "fedml[tensorflow,jax,mxnet]"

Installing with MPI support​

The MPI message exchanging mechanism is used to execute a local distributed simulation:

pip install "fedml[MPI]"

Installing from source​

First we need to clone the main repository and then install the FedML package using the setup.py script:

git clone https://github.com/FedML-AI/FedML.git && \
cd ./FedML/python && \
python setup.py install

If you want to run examples with TensorFlow, JAX or MXNet, you need to install the respective optional dependencies:

git clone https://github.com/FedML-AI/FedML.git && \
cd ./FedML/python && \
python install -e '.[tensorflow]'
python install -e '.[jax]'
python install -e '.[mxnet]'

If you need to install from a specific commit (usually for debug/dev purposes), please follow the commands below:

info

Remember to change the COMMIT_ID placeholder to the required id (look at the GitHub Commits page for all ids).

git clone https://github.com/FedML-AI/FedML.git && \
cd ./FedML && git checkout COMMIT_ID && \
cd python && \
python setup.py install

Troubleshooting​

If you meet any issues during installation or you have additional installation requirements, please post the issue at the GitHub Issue page.