Hands-On Artificial Intelligence for Beginners

Managing a TensorFlow model

TensorBoard is a tool built into TensorFlow that visualizes the values of your network's parameters over the coarse of the training process. It can also help to visualize data, and can run several exploratory analysis algorithms. It helps you explore the underlying TensorFlow graph in a simple, easy to use graphical interface. 

Once a TensorFlow model is trained, say the MNIST example defined precedingly, TensorFlow allows us to create something called a Model Summary. Summaries are, as you probably guessed, condensed versions of a model that contain the necessary key aspects that we need to use TensorBoard.

1. To create a summary, we first need to initialize the writer; we'll insert the line preceding just before ...