Some benchmarks for using Ray's implementation of joblib backend to distribute scikit-learn training.
How to run:
- run
ray up -y blog<#nodes>.yaml. - run
ray attach blog<#nodes>.yaml - copy the relevant <benchmark>.py to the remote cluster.
- run
RAY_ADDRESS=auto python <benchmark>.py