This hypergraph_nets library is heavily based off 'graph_nets', a framework for
general graph neural networks developed by Google Deepmind.

Their framework was implemented using Tensorflow and Sonnet: https://github.com/deepmind/graph_nets/
Our framework is implemented using PyTorch, and generalises graphs to hypergraphs.


Modules to rework:
- aggregators.py
- blocks.py
- broadcast.py
- hypergraphs.py


Modules done:
- models.py
- modules.py