Channel Pruning for Accelerating Very Deep Neural Networks | IEEE Conference Publication | IEEE Xplore