CONVOLUTIONAL NEURAL NETWORK PRUNING USING FILTER ATTENUATION

Filters are the essential elements in convolutional neural networks (CNNs). Filters are corresponded to the feature maps and form the main part of the computational and memory requirement for the CNN processing. In filter pruning methods, a filter with all of its components including channels and connections are removed. This can cause a harsh change in the network’s performance. Also, there are not any possibilities for the removed filters to come back to the network structure. To address this problem in this paper, a CNN pruning method based on filter attenuation is proposed in which weak filters are not directly removed. Instead, the weak filters are attenuated and gradually removed. Using attenuation approach, weak filters are not abruptly removed and there is a chance for these filters to return to CNN. The filter attenuation method is assessed using the VGG model on the Cifar10 image classification task. Simulation results show that the filter attenuation works with different pruning criteria and better results are observed in comparison with the conventional pruning methods.