layer_insertion_sensitivity_based

This code implements sensitivity based layer insertion for residual and feedforward neural network architectures for some activation functions. The application of the algorithm to image classification is implemented here, for using GD or minibatch SGD.