Compression-aware Training of Neural Networks using Frank-Wolfe

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zimmer, Max, Spiegel, Christoph, Pokutta, Sebastian
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916124204990464
author Zimmer, Max
Spiegel, Christoph
Pokutta, Sebastian
author_facet Zimmer, Max
Spiegel, Christoph
Pokutta, Sebastian
contents Many existing Neural Network pruning approaches rely on either retraining or inducing a strong bias in order to converge to a sparse solution throughout training. A third paradigm, 'compression-aware' training, aims to obtain state-of-the-art dense models that are robust to a wide range of compression ratios using a single dense training run while also avoiding retraining. We propose a framework centered around a versatile family of norm constraints and the Stochastic Frank-Wolfe (SFW) algorithm that encourage convergence to well-performing solutions while inducing robustness towards convolutional filter pruning and low-rank matrix decomposition. Our method is able to outperform existing compression-aware approaches and, in the case of low-rank matrix decomposition, it also requires significantly less computational resources than approaches based on nuclear-norm regularization. Our findings indicate that dynamically adjusting the learning rate of SFW, as suggested by Pokutta et al. (2020), is crucial for convergence and robustness of SFW-trained models and we establish a theoretical foundation for that practice.
format Preprint
id arxiv_https___arxiv_org_abs_2205_11921
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Compression-aware Training of Neural Networks using Frank-Wolfe
Zimmer, Max
Spiegel, Christoph
Pokutta, Sebastian
Machine Learning
Optimization and Control
Many existing Neural Network pruning approaches rely on either retraining or inducing a strong bias in order to converge to a sparse solution throughout training. A third paradigm, 'compression-aware' training, aims to obtain state-of-the-art dense models that are robust to a wide range of compression ratios using a single dense training run while also avoiding retraining. We propose a framework centered around a versatile family of norm constraints and the Stochastic Frank-Wolfe (SFW) algorithm that encourage convergence to well-performing solutions while inducing robustness towards convolutional filter pruning and low-rank matrix decomposition. Our method is able to outperform existing compression-aware approaches and, in the case of low-rank matrix decomposition, it also requires significantly less computational resources than approaches based on nuclear-norm regularization. Our findings indicate that dynamically adjusting the learning rate of SFW, as suggested by Pokutta et al. (2020), is crucial for convergence and robustness of SFW-trained models and we establish a theoretical foundation for that practice.
title Compression-aware Training of Neural Networks using Frank-Wolfe
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2205.11921