AutoML Tables release notes

This page documents production updates to AutoML Tables. You can periodically check this page for announcements about new or updated features, bug fixes, known issues, and deprecated functionality.

For a list of known issues for AutoML Tables, see Known issues.

You can see the latest product updates for all of Google Cloud on the Google Cloud page, browse and filter all release notes in the Google Cloud console, or programmatically access release notes in BigQuery.

To get the latest product updates delivered to you, add the URL of this page to your feed reader, or add the feed URL directly: https://cloud.google.com/feeds/automl-tables-release-notes.xml

January 03, 2024

The shutdown date for AutoML Tables has changed from Jan 23, 2024 to Mar 31, 2024.

January 23, 2023

AI Hub and the legacy versions of the following products are deprecated and will no longer be available on Google Cloud after January 17, 2024:

All the functionality of these legacy versions and new features are available on the Vertex AI platform. See Migrate to Vertex AI to learn how to migrate your resources.

June 01, 2020

April 03, 2020

Integration with VPC Service Controls is now in beta stage.

November 21, 2019

As part of AI Explanations, AutoML Tables now provides the option to show how each feature impacted an online prediction. This capability is called local feature importance, and is calculated using the Sampled Shapley method. Learn more.

November 18, 2019

Support for the European Union region, including the ability to configure AutoML Tables to store your data at rest and perform machine learning processing only in the European Union. Learn more.

Support for exporting AutoML Tables models to Cloud Storage, and then use Docker to make the model available for predictions. Learn more.

Support for using Stackdriver Logging to see final model hyperparameters as well as hyperparameters used during training trials. Learn more.

November 15, 2019

  • Support for up to 500 distinct values for Categorical target column.
  • Support for Precision at recall and Recall at precision optimization objectives for classification models. Learn more.

The AutoML Tables Python client library now includes additional methods that simplify using the AutoML API for common AutoML Tables tasks. Learn more.

July 23, 2019

Datasets smaller than 100,000 rows (and larger than the minimum size of 1,000 rows) are now fully supported.

June 28, 2019

Support for the "early stopping" feature. The model training process now stops by default when the search process is no longer finding better performing models. Early stopping can also be disabled.

June 12, 2019

  • Support for up to 100 distinct values for Categorical target column.
  • Support for BigQuery views.

April 29, 2019

Filename change for CSV output files for batch predictions; now tables_1.csv, tables_2.csv and so on. Learn more.

April 10, 2019

AutoML Tables Beta Release

March 19, 2019

You must deploy a model before you can request online predictions using that model. Once you deploy a model, it remains deployed until you undeploy it. You can deploy and undeploy models by using Google Cloud Platform Console or by using the Cloud AutoML. Learn more.

December 14, 2018

AutoML Tables EAP release

Only the us-central1 location is supported.