Takes a list of `layers`, describing the network architecture, and a value for the `mini_batch_size` to be used during training by stochastic gradient descent.
(self, layers, mini_batch_size)
source not stored for this graph (policy: none)
nothing calls this directly
no test coverage detected