super sparse convolution

sparse kernel (1) channel sparsity (1) computational flop (1) parameter reduction (1) parameter efficiency (1) convolutional neural network (1) computational efficiency (1) neural network optimization (1) model compression (1)