multiLayerPerceptron : Num a => AllAlgebra [ieva] a => TensorMonoid (ieva .cont) => Nat -> Tensor [ieva] a -\-> Tensor [ieva] a -> {default False _ : Bool} -> Tensor [ieva] a -\-> Tensor [ieva] aN-layer multi-layer perceptron with a specified activation function,
and flag for whether the last layer should have it