AWS has announced AutoGluon-Tabular, an open source AutoGluon capability that allows you to train machine learning models on tabular datasets from sources such as spreadsheets and database tables.
- AutoGluon is designed to be an easy-to-use and easy-to-extend AutoML toolkit, suitable for both machine learning beginners and experts.
- Using AutoGluon, you can train machine learning models for image classification, object detection, text classification, and tabular data prediction with little to no prior experience in machine learning.
- You can run AutoGluon anywhere—from your laptop or workstation, to a powerful Amazon Elastic Compute Cloud (Amazon EC2)—instance to take advantage of multiple cores and get results faster.
- AutoGluon-Tabular also includes novel techniques for multi-layer stack ensembling that significantly boosts model accuracy.
- Because AutoGluon is fully open source, transparent, and extensible, you have complete visibility into what it’s doing at every stage of the process and you can even bring in your own algorithms and use them with AutoGluon.
- AutoGluon-Tabular users only need to know how to use three Python functions: Dataset(), fit(), and predict().