Add support for Hyperparameter Tuning on Google Cloud AI Platform #15429

darshan-majithiya · 2021-04-19T11:35:57Z

The MLEngineStartTrainingJobOperator is modified in airflow/providers/google/cloud/operators/mlengine.py file.

The lines 1138-1142, 1192, 1214, 1279, and 1280 are added to the file.

These lines add support for Model training involving hyperparameter tuning on AI Platform.

boring-cyborg · 2021-04-19T11:36:00Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contribution Guide (https://github.com/apache/airflow/blob/master/CONTRIBUTING.rst)
Here are some useful points:

Pay attention to the quality of your code (flake8, pylint and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it’s a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

marcosmarxm · 2021-04-19T17:58:41Z

airflow/providers/google/cloud/operators/mlengine.py

@@ -1189,6 +1189,7 @@ def __init__(
        mode: str = 'PRODUCTION',
        labels: Optional[Dict[str, str]] = None,
        impersonation_chain: Optional[Union[str, Sequence[str]]] = None,
+        hyperparameters: Optional[Dict] = None,


Add parameter in class documentation and maybe a link to Google documentation can be helpful to users https://cloud.google.com/ai-platform/training/docs/using-hyperparameter-tuning and https://cloud.google.com/ai-platform/training/docs/reference/rest/v1/projects.jobs#HyperparameterSpec

:param hyperparameters Make a dictionary representing your HyperparameterSpec and add it to your training input. :type Dict

Updated and pushed the code

marcosmarxm · 2021-04-20T12:15:27Z

airflow/providers/google/cloud/operators/mlengine.py

+    :param hyperparameters: Optional HyperparameterSpec dictionary for hyperparameter tuning
+        For further reference, check these:
+        https://cloud.google.com/ai-platform/training/docs/using-hyperparameter-tuning
+        https://cloud.google.com/ai-platform/training/docs/reference/rest/v1/projects.jobs#HyperparameterSpec
+    :type hyperparameters: Dict


Take a look here

I think the way you wrote the documentation they wont be render as lists.
You can test locally using ./breeze build-docs -- --package-filter apache-airflow-providers-google to generate documentation.

I'm not able to run this locally. It exists without any error message.

I tried pushing the code to PR but the build throws an error for the 2nd URL line - https://cloud.google.com/ai-platform/training/docs/reference/rest/v1/projects.jobs#HyperparameterSpec. The line has 112 characters and the max allowed is 110. Will it be okay if I modify the /airflow/pylintrc file and update the max-line-length value to 112? Or is there any other workaround?

Did you enable pre-commit to validate pylint/flake etc?

I'm not able to run pre-commit on my system. It's throwing permission error. Been trying to find a solution for a while, but haven't figured it yet.

Meanwhile, I've updated and pushed the documentation as per the demo in the link you had shared. Please check it out.

github-actions · 2021-04-21T06:45:19Z

The Workflow run is cancelling this PR. Building images for the PR has failed. Follow the workflow link to check the reason.

xinbinhuang · 2021-04-21T15:57:23Z

Hi @darshan-majithiya,

Thanks for the PR! Can you also add a test case to the file https://github.com/apache/airflow/blob/master/tests/providers/google/cloud/operators/test_mlengine.py ?

leahecole · 2021-04-23T21:30:10Z

+1 to the request for test case, and also consider adding an example to https://github.com/apache/airflow/blob/master/airflow/providers/google/cloud/example_dags/example_mlengine.py

Thanks for your contribution! :)

github-actions · 2021-07-23T00:04:59Z

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in 5 days if no further activity occurs. Thank you for your contributions.

mik-laj · 2021-07-23T00:51:53Z

airflow/providers/google/cloud/operators/mlengine.py

@@ -1274,6 +1280,9 @@ def execute(self, context):
        if self._service_account:
            training_request['trainingInput']['serviceAccount'] = self._service_account

+        if self._hyperparameters:
+            training_request['trainingInput']['hyperparameters'] = self._hyperparameters


Can you add tests to avoid regression?

Adding support for Hyperparameter Tuning on Google Cloud AI Platform

c77eb1f

darshan-majithiya requested a review from turbaszek as a code owner April 19, 2021 11:35

boring-cyborg bot added area:providers provider:google Google (including GCP) related issues labels Apr 19, 2021

darshan-majithiya added 2 commits April 19, 2021 17:37

Fix syntax on line 1192

a97e8d2

Remove trailing white space

10f29f8

marcosmarxm suggested changes Apr 19, 2021

View reviewed changes

darshan-majithiya force-pushed the gcp-ai-platform-hyperparameter-tuning branch from 66d839e to 10f29f8 Compare April 20, 2021 07:59

Update Class Documentation

7c19baa

marcosmarxm suggested changes Apr 20, 2021

View reviewed changes

darshan-majithiya force-pushed the gcp-ai-platform-hyperparameter-tuning branch from c91463d to 7c19baa Compare April 21, 2021 06:20

darshan-majithiya changed the title ~~Adding support for Hyperparameter Tuning on Google Cloud AI Platform~~ Add support for Hyperparameter Tuning on Google Cloud AI Platform Apr 21, 2021

darshan-majithiya force-pushed the gcp-ai-platform-hyperparameter-tuning branch from ddf814d to 7c19baa Compare April 21, 2021 07:16

Update MLEngineStartTrainingJobOperator Class Documentation

c85762a

github-actions bot added the stale Stale PRs per the .github/workflows/stale.yml policy file label Jul 23, 2021

mik-laj reviewed Jul 23, 2021

View reviewed changes

github-actions bot closed this Jul 30, 2021

leahecole mentioned this pull request Jul 30, 2021

Add support for hyperparameter tuning on GCP Cloud AI #17348

Closed

subkanthi mentioned this pull request Aug 23, 2021

Gcp ai hyperparameter tuning #17790

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add support for Hyperparameter Tuning on Google Cloud AI Platform #15429

Add support for Hyperparameter Tuning on Google Cloud AI Platform #15429

darshan-majithiya commented Apr 19, 2021 •

edited

Loading

boring-cyborg bot commented Apr 19, 2021

marcosmarxm Apr 19, 2021 •

edited

Loading

darshan-majithiya Apr 20, 2021

marcosmarxm Apr 20, 2021

darshan-majithiya Apr 21, 2021

marcosmarxm Apr 21, 2021

darshan-majithiya Apr 25, 2021 •

edited

Loading

github-actions bot commented Apr 21, 2021

xinbinhuang commented Apr 21, 2021

leahecole commented Apr 23, 2021

github-actions bot commented Jul 23, 2021

mik-laj Jul 23, 2021

Add support for Hyperparameter Tuning on Google Cloud AI Platform #15429

Add support for Hyperparameter Tuning on Google Cloud AI Platform #15429

Conversation

darshan-majithiya commented Apr 19, 2021 • edited Loading

boring-cyborg bot commented Apr 19, 2021

marcosmarxm Apr 19, 2021 • edited Loading

Choose a reason for hiding this comment

darshan-majithiya Apr 20, 2021

Choose a reason for hiding this comment

marcosmarxm Apr 20, 2021

Choose a reason for hiding this comment

darshan-majithiya Apr 21, 2021

Choose a reason for hiding this comment

marcosmarxm Apr 21, 2021

Choose a reason for hiding this comment

darshan-majithiya Apr 25, 2021 • edited Loading

Choose a reason for hiding this comment

github-actions bot commented Apr 21, 2021

xinbinhuang commented Apr 21, 2021

leahecole commented Apr 23, 2021

github-actions bot commented Jul 23, 2021

mik-laj Jul 23, 2021

Choose a reason for hiding this comment

darshan-majithiya commented Apr 19, 2021 •

edited

Loading

marcosmarxm Apr 19, 2021 •

edited

Loading

darshan-majithiya Apr 25, 2021 •

edited

Loading