README.md

XGBoost Plugin for Federated Learning

This folder contains the plugin for federated learning. Follow these steps to build and test it.

Note. Building XGBoost with -DPLUGIN_FEDERATED=ON flag will automatically download the gRPC source code and build it, along with its dependencies. This will increase compilation time. If you already have gRPC installed on your system, pass additional flag -DUSE_GRPC_FROM_SYSTEM=ON to speed up the build.

Build the Plugin

# Under xgboost source tree.
mkdir build
cd build
cmake .. -GNinja\
 -DPLUGIN_FEDERATED=ON\
 -DBUILD_WITH_CUDA_CUB=ON\
 -DUSE_CUDA=ON\
 -DUSE_NCCL=ON
ninja
cd ../python-package
pip install -e .  # or equivalently python setup.py develop

Test Federated XGBoost

# Under xgboost source tree.
cd tests/distributed
# This tests both CPU training (`hist`) and GPU training (`gpu_hist`).
./runtests-federated.sh