This lab provides hands-on practice serving machine learning predictions in batch and real time.
You’ll begin by creating a prediction service to call a trained model that has been deployed in Google Cloud.
From there, you’ll practice running a Dataflow job where the prediction service reads in batches from a CSV file.
CSV
And, finally, you’ll learn how to run a streaming Dataflow pipeline to read requests in real time from Cloud Pub/Sub,
and also write predictions to a BigQuery table.