IPPRO: Importance-based Pruning with PRojective Offset for Magnitude-indifferent Structural Pruning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jung, Jaeheun, Lee, Jaehyuk, Lee, Yeajin, Lee, Donghun
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911068972908544
author Jung, Jaeheun
Lee, Jaehyuk
Lee, Yeajin
Lee, Donghun
author_facet Jung, Jaeheun
Lee, Jaehyuk
Lee, Yeajin
Lee, Donghun
contents With the growth of demand on neural network compression methods, the structured pruning methods including importance-based approach are actively studied. The magnitude importance and many correlated modern importance criteria often limit the capacity of pruning decision, since the filters with larger magnitudes are not likely to be pruned if the smaller one didn't, even if it is redundant. In this paper, we propose a novel pruning strategy to challenge this dominating effect of magnitude and provide fair chance to each filter to be pruned, by placing it on projective space. After that, we observe the gradient descent movement whether the filters move toward the origin or not, to measure how the filter is likely to be pruned. This measurement is used to construct PROscore, a novel importance score for IPPRO, a novel importance-based structured pruning with magnitude-indifference. Our evaluation results shows that the proposed importance criteria using the projective space achieves near-lossless pruning by reducing the performance drop in pruning, with promising performance after the finetuning. Our work debunks the ``size-matters'' myth in pruning and expands the frontier of importance-based pruning both theoretically and empirically.
format Preprint
id arxiv_https___arxiv_org_abs_2507_14171
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle IPPRO: Importance-based Pruning with PRojective Offset for Magnitude-indifferent Structural Pruning
Jung, Jaeheun
Lee, Jaehyuk
Lee, Yeajin
Lee, Donghun
Machine Learning
Artificial Intelligence
With the growth of demand on neural network compression methods, the structured pruning methods including importance-based approach are actively studied. The magnitude importance and many correlated modern importance criteria often limit the capacity of pruning decision, since the filters with larger magnitudes are not likely to be pruned if the smaller one didn't, even if it is redundant. In this paper, we propose a novel pruning strategy to challenge this dominating effect of magnitude and provide fair chance to each filter to be pruned, by placing it on projective space. After that, we observe the gradient descent movement whether the filters move toward the origin or not, to measure how the filter is likely to be pruned. This measurement is used to construct PROscore, a novel importance score for IPPRO, a novel importance-based structured pruning with magnitude-indifference. Our evaluation results shows that the proposed importance criteria using the projective space achieves near-lossless pruning by reducing the performance drop in pruning, with promising performance after the finetuning. Our work debunks the ``size-matters'' myth in pruning and expands the frontier of importance-based pruning both theoretically and empirically.
title IPPRO: Importance-based Pruning with PRojective Offset for Magnitude-indifferent Structural Pruning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2507.14171