Salvato in:
Dettagli Bibliografici
Autori principali: Gharatappeh, Soheil, Sekeh, Salimeh Yasaei
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2501.15592
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910801784209408
author Gharatappeh, Soheil
Sekeh, Salimeh Yasaei
author_facet Gharatappeh, Soheil
Sekeh, Salimeh Yasaei
contents Iterative magnitude pruning methods (IMPs), proven to be successful in reducing the number of insignificant nodes in over-parameterized deep neural networks (DNNs), have been getting an enormous amount of attention with the rapid deployment of DNNs into cutting-edge technologies with computation and memory constraints. Despite IMPs popularity in pruning networks, a fundamental limitation of existing IMP algorithms is the significant training time required for each pruning iteration. Our paper introduces a novel \textit{stopping criterion} for IMPs that monitors information and gradient flows between networks layers and minimizes the training time. Information Consistent Pruning (\ourmethod{}) eliminates the need to retrain the network to its original performance during intermediate steps while maintaining overall performance at the end of the pruning process. Through our experiments, we demonstrate that our algorithm is more efficient than current IMPs across multiple dataset-DNN combinations. We also provide theoretical insights into the core idea of our algorithm alongside mathematical explanations of flow-based IMP. Our code is available at \url{https://github.com/Sekeh-Lab/InfCoP}.
format Preprint
id arxiv_https___arxiv_org_abs_2501_15592
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Information Consistent Pruning: How to Efficiently Search for Sparse Networks?
Gharatappeh, Soheil
Sekeh, Salimeh Yasaei
Machine Learning
Information Theory
Neural and Evolutionary Computing
Iterative magnitude pruning methods (IMPs), proven to be successful in reducing the number of insignificant nodes in over-parameterized deep neural networks (DNNs), have been getting an enormous amount of attention with the rapid deployment of DNNs into cutting-edge technologies with computation and memory constraints. Despite IMPs popularity in pruning networks, a fundamental limitation of existing IMP algorithms is the significant training time required for each pruning iteration. Our paper introduces a novel \textit{stopping criterion} for IMPs that monitors information and gradient flows between networks layers and minimizes the training time. Information Consistent Pruning (\ourmethod{}) eliminates the need to retrain the network to its original performance during intermediate steps while maintaining overall performance at the end of the pruning process. Through our experiments, we demonstrate that our algorithm is more efficient than current IMPs across multiple dataset-DNN combinations. We also provide theoretical insights into the core idea of our algorithm alongside mathematical explanations of flow-based IMP. Our code is available at \url{https://github.com/Sekeh-Lab/InfCoP}.
title Information Consistent Pruning: How to Efficiently Search for Sparse Networks?
topic Machine Learning
Information Theory
Neural and Evolutionary Computing
url https://arxiv.org/abs/2501.15592