Implementation of a hierarchical Gaussian process as in Hensman et. al. using GPFlow2. Fundamentally this this model is just a new kernel that can then be used in any of the GPModule inherited objects in GPFlow. The kernel can just be import from kernel.py and a demo notebook is also provided. It is worth running this in a virtual environment using the dependencies listed in requirements.txt as TensorFlow changes quite significantly from version-to-version.