Training a large neural network and then making it small by pruning it's weights/neurons
I am uploading the final results after pruning the network. Firstly, "weights pruning" . Secondly, "Unit/Neuron pruning" . Plots of "percent sparsity Vs accuracy" for both techniques.