Pruning Filters while Training for Efficiently Optimizing Deep Learning Networks