Challenge 4: Automagic training with pipelines

< Previous Challenge - Home - Next Challenge >

Introduction

The previous challenge introduced the concept of build pipelines. But there are different types of pipelines, and this task is getting started with Vertex AI pipelines for continuous training.

Description

If you’ve successfully completed the previous challenge, your training code has been packaged and can be run from a pipeline.

The provided project has a pipeline.py file that can generate a pipeline definition. Run that to generate a pipeline definition file (JSON). Use the generated pipeline definition file to create a new Pipeline Run through the GCP Console. You’ll need to fill in some parameters (you can look up the Python package location). Do not set/override the endpoint and monitoring_job parameters (keep the default values).

Once the pipeline is triggered, it will take ~10 minutes to complete.

Success Criteria

1. There’s at least one successful Vertex AI pipeline run that has generated a Managed Model in the Model Registry
2. No code change is needed for this challenge

Tips

• You have already created a bucket, you can use that as the pipeline root (optionally add pipelines folder in it).
• Make sure that you’re consistent in your region selection.
• You can either upload the pipeline definition from a local machine, or put it on GCS and refer to its location.
• Make sure that you’re running the module trainer.pipeline in the virtual environment you have created as part of the first challenge.
• If you’re in doubt about the parameters, remember to Use the Force and read the Source ;)

Learning Resources

• Running Python modules from the command line
• Running Vertex AI Pipelines on the console