See “A poor man’s orchestration of predictive models, or do it yourself” for further details.
In order to build a Docker image and upload it to Container Registry, run
make build
In order to (1) start a Compute Engine instance for training a model, (2) wait until completion, and (3) ensure that everything went well, run
make training-start training-wait training-check
In order to (1) start a Compute Engine instance for applying a trained model, (2) wait until completion, and (3) ensure that everything went well, run
make application-start application-wait application-check
In order to schedule training and application using Airflow, run
git clone https://github.com/chain-rule/example-prediction-service.git "${AIRFLOW_HOME}/dags"