Now, we will visualize the training result with TensorBoard.

First, you need to change the command-line directory to the folder that contains the checkpoints. In our case, it is the train_dir parameter in the previous command, D:datasetsdiabeticcheckpoints. Then, you should run the following command:

tensorboard -logdir .

Here is some output when we run TensorBoard for our network:

The preceding image shows the nodes containing the RMS prop optimizer for the training network and some logits that it contains for the output of DR classification. The next screenshot shows the images coming as input, along with their preprocessing and modifications:

In this screenshot, you can see the graph showing the network output during training:

This screenshot depicts the total raw loss of the network during training: