Hands-On Machine Learning on Google Cloud Platform

The __init__ file

For Cloud ML engine to create a package for the module we are building, it needs to create a package for the module. For the package to be created, it needs to create the __init__.py file in the trainer folder.

For that, we would run the following code:

touch trainer/__init__.py

Now that the whole setup is ready, we run the job as follows:

export BUCKET_NAME=my-mnist-bucketexport JOB_NAME="mnist_mlp_hpt_train_$(date +%Y%m%d_%H%M%S)"export JOB_DIR=gs://$BUCKET_NAME/$JOB_NAMEexport REGION=us-east1export HPTUNING_CONFIG=hptune.yamlgcloud ml-engine jobs submit training $JOB_NAME \ --job-dir $JOB_DIR \ --runtime-version 1.6 \ --config $HPTUNING_CONFIG \ --module-name trainer.mnist_mlp_lr_numsteps \ --package-path ./trainer \ ...