Deep neural networks often have millions of parameters. This can hinder their deployment to low-end devices, not only due to high memory requirements but also because of increased latency at inference. We propose a novel model compression method that generates a sparse trained model without additional overhead: by allowing (i) dynamic allocation of the sparsity pattern and (ii) incorporating feedback signal to reactivate prematurely pruned weights we obtain a performant sparse model in one single training pass (retraining is not needed, but can further improve the performance). We evaluate our method on CIFAR-10 and ImageNet, and show that the obtained sparse models can reach the state-of-the-art performance of dense models. Moreover, their performance surpasses that of models generated by all previously proposed pruning schemes.
@article{arxiv.2006.07253,
title = {Dynamic Model Pruning with Feedback},
author = {Tao Lin and Sebastian U. Stich and Luis Barba and Daniil Dmitriev and Martin Jaggi},
journal= {arXiv preprint arXiv:2006.07253},
year = {2020}
}