An Efficient End-to-End Channel Level Pruning Method for Deep Neural Networks Compression | IEEE Conference Publication | IEEE Xplore