The tfruns package is a set of utilities for managing different training runs for deep learning models. It can be used as a framework to build multiple deep learning models using different hyper-parameters. It can track the hyper-parameters, metrics, output, and source code for every training run and allows you to compare the best models so that you can see the differences between the training runs. This makes hyper-parameter tuning much easier and can be used with any tfestimator model or Keras model. For more details, go to https://tensorflow.rstudio.com/tools/tfruns/articles/overview.html.
The following code is in the Chapter8/hyperparams.R folder and also uses the script we used in the TensorFlow estimators section ...
