arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.01290cs.LGcs.CV
收藏

稀疏神经网络的周期剪枝

Cyclical Pruning for Sparse Neural Networks

Idiap Research Institute(IDIAP研究所);EPFL(洛桑联邦理工学院);Qualcomm AI Research(高通人工智能研究院)

Suraj Srinivas, Andrey Kuzmin, Markus Nagel, Mart van Baalen, Andrii Skliar, Tijmen Blankevoort

摘要:

用于修剪神经网络权重的当前方法迭代地对模型权重应用基于幅度的修剪,并重新训练所得到的模型以恢复丢失的准确性。在这项工作中,我们表明,这种策略不允许错误修剪的权重的恢复。为了使权重恢复,我们提出了一个简单的策略称为\textit{cyclical pruning},它需要定期修剪时间表,并允许在一个周期中错误修剪的权重在随后的恢复。在线性模型和大规模深度神经网络上的实验结果表明,循环剪枝优于现有的剪枝算法,特别是在高稀疏率下。我们的方法很容易调整,可以很容易地纳入现有的修剪管道,以提高性能。

英文摘要:

Current methods for pruning neural network weights iteratively apply magnitude-based pruning on the model weights and re-train the resulting model to recover lost accuracy. In this work, we show that such strategies do not allow for the recovery of erroneously pruned weights. To enable weight recovery, we propose a simple strategy called \textit{cyclical pruning} which requires the pruning schedule to be periodic and allows for weights pruned erroneously in one cycle to recover in subsequent ones. Experimental results on both linear models and large-scale deep neural networks show that cyclical pruning outperforms existing pruning algorithms, especially at high sparsity ratios. Our approach is easy to tune and can be readily incorporated into existing pruning pipelines to boost performance.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑