Global Sparse Momentum SGD for Pruning Very Deep Neural Networks

TL;DR

This paper introduces a global sparse momentum SGD method for pruning deep neural networks, enabling automatic, end-to-end model compression without extensive manual tuning or post-pruning retraining.

Contribution

The proposed method achieves automatic layer-wise sparsity, simplifies pruning process, and improves the discovery of effective subnetworks compared to prior techniques.

Findings

Automatic global sparsity ratios for all layers.

No need for post-pruning retraining.

Better identification of winning subnetworks.

Abstract

Deep Neural Network (DNN) is powerful but computationally expensive and memory intensive, thus impeding its practical usage on resource-constrained front-end devices. DNN pruning is an approach for deep model compression, which aims at eliminating some parameters with tolerable performance degradation. In this paper, we propose a novel momentum-SGD-based optimization method to reduce the network complexity by on-the-fly pruning. Concretely, given a global compression ratio, we categorize all the parameters into two parts at each training iteration which are updated using different rules. In this way, we gradually zero out the redundant parameters, as we update them using only the ordinary weight decay but no gradients derived from the objective function. As a departure from prior methods that require heavy human works to tune the layer-wise sparsity ratios, prune by solving complicated non-differentiable problems or finetune the model after pruning, our method is characterized by 1) global compression that automatically finds the appropriate per-layer sparsity ratios; 2) end-to-end training; 3) no need for a time-consuming re-training process after pruning; and 4) superior capability to find better winning tickets which have won the initialization lottery.