Summary
In this chapter, we learned how to use Keras Tuner in Google Cloud AI Platform. We learned how to run the hyperparameter search, and we learned how to train a model with the best hyperparameter configuration. We have also seen that in a typical Keras style, integrating Keras Tuner into our existing model training workflow is very easy, especially with the simple treatment of hyperparameters as just arrays of a certain data type. This really opens up the choices for hyperparameters, and we do not need to implement the search logic or complicated conditional loops to keep track of the results.
In the next chapter, we will see the latest model optimization techniques that reduce the model size. As a result, our model can be leaner and more compact.