Professional Cloud Architect – Google Cloud Certification Guide

Submitting a training job

Before you run a training job in the ML Engine, you can actually run the job locally on your machine, or in Cloud Shell using the gcloud ml-engine local command.

This allows you to adjust the job quickly and not have to pay for ML Engine resources.

Exam tip: Remember that you can use ML Engine locally when you develop your model to save time and costs.

Once you are ready to execute training with large datasets, use Cloud ML Engine.

To submit a training model job to ML Engine, you can use either Google Cloud Console or the gcloud ml-engine jobs submit command. Let's have a look at the syntax for a job submission:

 gcloud ml-engine jobs submit training $JOB_NAME \         --scale-tier basic \ --package-path $PACKAGE_PATH ...