Introducing the AutoGluon library
AutoGluon is a Python library developed by AWS and open sourced at their annual re:Invent conference, in 2019. The primary design goal behind AutoGluon is similar to SageMaker's Autopilot module – to resolve all the complexities and challenges an ML practitioner faces in a typical ML process and resolve these with a single Python library. In essence, AutoGluon empowers the ML practitioner to organize their training data and apply several ML approaches to generate an optimized model, all with just a few lines of code.
AutoGluon overcomes some of the limitations that AutoPilot has in that it can address the more complex ML use cases that involve compound types of data, such as cluttered text data and images. Of course, AutoGluon also works with tabular data. AutoGluon accomplishes this by creating separate predictors for each data type and, hence, each type of ML use case that the data type supports. For example, AutoGluon includes the...
