[tehsauce]

The concept of "layers" is not in fact enforced by pytorch or tensorflow at all. This tutorial is a really nice overview of the levels of abstraction available in pytorch https://pytorch.org/tutorials/beginner/nn_tutorial.html