Hands-on exercise – training a TensorFlow model

In this exercise, you will learn how to install the TensorFlow library in your local Jupyter environment and build and train a simple neural network model. Launch a Jupyter notebook that you have previously installed on your machine. If you don't remember how to do this, visit the Hands-on lab section of Chapter 3, Machine Learning Algorithms.

Once the Jupyter notebook is running, create a new folder by selecting the New dropdown and then Folder. Rename the folder TensorFlowLab. Open the TensorFlowLab folder, create a new notebook inside this folder, and rename the notebook Tensorflow-lab1.ipynb. Now, let's get started:

1. Inside the first cell, run the following code to install TensorFlow:

   ! pip3 install --upgrade tensorflow

2. Now, we must import the library and load the sample training data. We will use the built-in fashion_mnist dataset that comes with the keras library to do so. Next, we must load the data...