SR-init: An interpretable layer pruning method
This addresses model compression for efficient deployment, but it is incremental as it builds on existing pruning methods by adding interpretability.
The paper tackles the challenge of deploying deep neural networks on resource-constrained devices by proposing SR-init, an interpretable layer pruning method that reduces parameters by up to 63.98% on CIFAR-10 with only a 0.56% drop in top-1 accuracy.
Despite the popularization of deep neural networks (DNNs) in many fields, it is still challenging to deploy state-of-the-art models to resource-constrained devices due to high computational overhead. Model pruning provides a feasible solution to the aforementioned challenges. However, the interpretation of existing pruning criteria is always overlooked. To counter this issue, we propose a novel layer pruning method by exploring the Stochastic Re-initialization. Our SR-init method is inspired by the discovery that the accuracy drop due to stochastic re-initialization of layer parameters differs in various layers. On the basis of this observation, we come up with a layer pruning criterion, i.e., those layers that are not sensitive to stochastic re-initialization (low accuracy drop) produce less contribution to the model and could be pruned with acceptable loss. Afterward, we experimentally verify the interpretability of SR-init via feature visualization. The visual explanation demonstrates that SR-init is theoretically feasible, thus we compare it with state-of-the-art methods to further evaluate its practicability. As for ResNet56 on CIFAR-10 and CIFAR-100, SR-init achieves a great reduction in parameters (63.98% and 37.71%) with an ignorable drop in top-1 accuracy (-0.56% and 0.8%). With ResNet50 on ImageNet, we achieve a 15.59% FLOPs reduction by removing 39.29% of the parameters, with only a drop of 0.6% in top-1 accuracy. Our code is available at https://github.com/huitang-zjut/SR-init.