Installing FedML¶
FedML supports Linux, MacOS, Windows, and Android.
FedML Source Code Repository¶
Installing with pip¶
pip install fedml
The default machine learning engine is PyTorch. FedML also supports TensorFlow, Jax, and MXNet.
You can install related engine as follows:
pip install "fedml[tensorflow]"
pip install "fedml[jax]"
pip install "fedml[mxnet]"
The above commands work properly in Linux environment. For Windows/Mac OS (Intel)/Mac OS (M1), you may need to follow TensorFlow/Jax/MXNet official guidance to fix related installation issues.
Installing FedML with Anaconda¶
conda create --name fedml python
conda activate fedml
conda install --name fedml pip
pip install fedml
(Note: please use python 3.7 if you met any compatability issues. We will support 3.8, 3.9, 3.10 systematically in the next iteration.)
After installation, please use “pip list | grep fedml” to check whether fedml is installed.
Installing FedML from Debugging and Editable Mode¶
cd python
pip install -e ./
Installing FedML from Source¶
git clone https://github.com/FedML-AI/FedML.git && \
cd ./FedML/python && \
python setup.py install
If you want to run examples with tensforflow, jax or mxnet, you need to install 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]'
(Notes: Tensorflow example located in tf_mqtt_s3_fedavg_mnist_lr_example directory, Jax example location in jax_haiku_mqtt_s3_fedavg_mnist_lr_example directory)
If you need to install from a specific commit (normally used for the debugging/development phase), please follow commands below:
git clone https://github.com/FedML-AI/FedML.git && \
cd ./FedML && git checkout e798061d62560b03e049d514e7cc8f1a753fde6b && \
cd python && \
python setup.py install
Please change the above commit id to your own (you can find it at https://github.com/FedML-AI/FedML/commits/master)
Running FedML in Docker (Recommended)¶
FedML Docker Hub: https://hub.docker.com/repository/docker/fedml/fedml
We recommend using FedML in the Docker environment as it circumvents complex and tedious installation debugging. Currently, we maintain docker images for two settings:
For Linux servers with x86_64 architecture
Please refer to the following commands and remember to change WORKSPACE to your own.
(1) Pull the Docker image and prepare the docker environment
FEDML_DOCKER_IMAGE=fedml/fedml:latest-torch1.12.1-cuda11.3-cudnn8-devel
docker pull FEDML_DOCKER_IMAGE
# if you want to use GPUs in your host OS, please follow this link: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker
sudo apt-get update
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker
sudo chmod 777 /var/run/docker.sock
(2) Run Docker with interactive mode
FEDML_DOCKER_IMAGE=fedml/fedml:latest-torch1.12.1-cuda11.3-cudnn8-devel
WORKSPACE=/home/chaoyanghe/sourcecode/FedML_startup/FedML
docker run -t -i -v $WORKSPACE:$WORKSPACE --shm-size=64g --ulimit nofile=65535 --ulimit memlock=-1 --privileged \
--env FEDML_NODE_INDEX=0 \
--env WORKSPACE=$WORKSPACE \
--env FEDML_NUM_NODES=1 \
--env FEDML_MAIN_NODE_INDEX=0 \
--env FEDML_RUN_ID=0 \
--env FEDML_MAIN_NODE_PRIVATE_IPV4_ADDRESS=127.0.0.1 \
--gpus all \
-u fedml --net=host \
$FEDML_DOCKER_IMAGE \
/bin/bash
You should now see a prompt that looks something like:
fedml@ChaoyangHe-GPU-RTX2080Tix4:/$
fedml@ChaoyangHe-GPU-RTX2080Tix4:/$ cd $WORKSPACE
fedml@ChaoyangHe-GPU-RTX2080Tix4:/home/chaoyanghe/sourcecode/FedML_startup/FedML$
If you want to debug in Docker container, please follow these commands
cd python
# You need sudo permission to install your debugging package in editable mode
(-e means link the package to the original location, basically meaning any changes to the original package would reflect directly in your environment)
sudo pip install -e ./
(3) Run Docker with multiple commands to launch your project immediately
Here is an example of running federated learning with MNIST dataset and Logistic Regression model.
WORKSPACE=/home/chaoyanghe/sourcecode/FedML_startup/FedML
docker run -t -i -v $WORKSPACE:$WORKSPACE --shm-size=64g --ulimit nofile=65535 --ulimit memlock=-1 --privileged \
--env FEDML_NODE_INDEX=0 \
--env WORKSPACE=$WORKSPACE \
--env FEDML_NUM_NODES=1 \
--env FEDML_MAIN_NODE_INDEX=0 \
--env FEDML_RUN_ID=0 \
--env FEDML_MAIN_NODE_PRIVATE_IPV4_ADDRESS=127.0.0.1 \
-u fedml --net=host \
--gpus all \
$FEDML_DOCKER_IMAGE \
/bin/bash -c `cd $WORKSPACE/python/examples/simulation/mpi_torch_fedavg_mnist_lr_example; sh run_one_line_example.sh`
(4) Run Docker with bootstrap.sh and entry.sh
For advanced usage, you may need to install additional python packages or set some additional environments for your project.
In this case, we recommend that you specify the bootstrap.sh for the additional package installation and environment settings, and
entry.sh, where you launch your main program. Here is an example of running the same task in (3).
——-boostrap.sh———-
#!/bin/bash
echo "This is bootstrap script. You can use it to customize your additional installation and set some environment variables"
# here we upgrade fedml to the latest version.
pip install --upgrade fedml
——-entry.sh———-
#!/bin/bash
echo "This is entry script where you launch your main program."
cd $WORKSPACE/python/examples/simulation/mpi_torch_fedavg_mnist_lr_example
sh run_one_line_example.sh
WORKSPACE=/home/chaoyanghe/sourcecode/FedML_startup/FedML
docker run -t -i -v $WORKSPACE:$WORKSPACE --shm-size=64g --ulimit nofile=65535 --ulimit memlock=-1 --privileged \
--env FEDML_NODE_INDEX=0 \
--env WORKSPACE=$WORKSPACE \
--env FEDML_NUM_NODES=1 \
--env FEDML_MAIN_NODE_INDEX=0 \
--env FEDML_RUN_ID=0 \
--env FEDML_MAIN_NODE_PRIVATE_IPV4_ADDRESS=127.0.0.1 \
--env FEDML_BATCH_BOOTSTRAP=$WORKSPACE/python/scripts/docker/bootstrap.sh \
--env FEDML_BATCH_ENTRY_SCRIPT=$WORKSPACE/python/scripts/docker/entry.sh \
--gpus all \
-u fedml --net=host \
$FEDML_DOCKER_IMAGE
(5) Run the interpreter in PyCharm or Visual Studio using Docker environment
PyCharm
https://www.jetbrains.com/help/pycharm/using-docker-as-a-remote-interpreter.html#summary
Visual Studio
https://code.visualstudio.com/docs/remote/containers
(6) Other useful commands
# docker rm $(docker ps -aq)
docker container kill $(docker ps -q)
Guidance for Windows Users¶
Please follow instructions at Windows Installation
Guidance for Raspberry Pi Users¶
Please follow instructions at Raspberry Pi Installation
Guidance for NVIDIA Jetson Devices¶
Please follow instructions at NVIDIA Jetson Device Installation
Testing if the installation succeeded¶
If the installation is successful, you will not see any issue when run import fedml.
(mnn37) chaoyanghe@Chaoyangs-MBP FedML-refactor % python
Python 3.7.7 (default, Mar 26 2020, 10:32:53)
[Clang 4.0.1 (tags/RELEASE_401/final)] :: Anaconda, Inc. on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import fedml
>>>
Installing FedML Android SDK/APP¶
Please follow the instructions at https://github.com/FedML-AI/FedML/java/README.md
Troubleshooting¶
If you met any issues during installation or have additional installation requirements, please post issues at https://github.com/FedML-AI/FedML/issues