AutoML

Some of the services we discussed in the previous sections use completely pre-trained models, and others allow you to bring your own data to either train or up-train a model based on your data. Pre-trained models are trained on datasets provided by Google or other sources, and the term, up-train, refers to augmenting a pre-trained model with additional data. If you want to create more customized use cases than those supported by the high-level API services, you may want to train your own models. Vertex AI, which we will describe later in this chapter, provides a plethora of tools for implementing every step in the model development process. However, before we get to the level of customizing every step in the process, one way in which you can easily start getting inferences from ML models that are trained on your data in Google Cloud is to use AutoML, which enables developers with limited ML expertise to train models specific to their business needs in as little as a few minutes...