Tensorflow basics on Jupyter Notebook + Virtualenv

Oct 31 2019

In order to set up a Jupyter Notebook session with a virtual environment, first create your virtual environment and activate it. See link for a How To:
https://uoa-eresearch.github.io/eresearch-cookbook/recipe/2014/11/26/python-virtual-env/

Then within the activated virtual environment run
pip install ipython
and
pip install ipykernel.
This will allow you to install a new kernal to appear in your jupyter notebook:
ipython kernel install --user --name=deep_learning_venv
Now you can select deep_learning_venv in the New Notebook dropdown.

To use Tensorflow you can then treat the virtual environment as usual. i.e pip install tensorflow, then you can import tensorflow as tf in the actual notebook cells.
I also installed matplotlib in order to view some of the neural network training data as an image directly in jupyter notebook.
E.g:

import matplotlib.pyplot as plt
plt.imshow(x_train[0], cmap=plt.cm.binary)
plt.show()

You can now start training a model.