This page lists known issues with AutoML Tables, along with ways you can avoid or recover from these issues.
Datasets with less than 100,000 rows can result in decreased model quality.
Datasets smaller than 100,000 rows might not get the full benefit of our more advanced modeling pipelines, leading to lower model quality than models trained from larger datasets.
To avoid this issue, include at least 100,000 rows of data in your dataset.
High latency for online prediction requests after deploying the model
After you deploy your model, the first online prediction requests show increased latency.
To avoid this issue, send a few prediction requests to warm up the model before using it in production.
Batch prediction requests with only one feature column fail.
You must provide at least 2 feature columns with batch predictions.
Batch prediction destination project access problem shown as internal error
When you use BigQuery as the destination for your batch prediction results, you specify the destination project. If you specify a project that does not exist, or for which you do not have sufficient permissions to create a dataset, AutoML Tables returns "Internal error".
If you get an internal error returned from a batch prediction request, confirm that you have provided the correct project name, and that you have permission to create a dataset in that project.
Using Google Cloud Platform Console with AutoML Tables
User experience with Microsoft Edge and Microsoft Internet Explorer browsers might be suboptimal.
Microsoft Edge and Microsoft Internet Explorer do not support all features of AutoML Tables. If you are having problems, try Google Chrome, Safari, or Firefox.
The following issues were listed on this page, but are no longer affecting AutoML Tables.
Training for longer than needed can degrade model quality.
Training for longer than needed no longer degrades model quality. Also, AutoML Tables automatically stops training the model when it detects that model quality is no longer improving.