Computer Science > Computer Vision and Pattern Recognition
[Submitted on 10 Mar 2020 (v1), revised 12 Mar 2020 (this version, v2), latest version 10 Sep 2020 (v5)]
Title:Channel Pruning via Optimal Thresholding
View PDFAbstract:Structured pruning, especially channel pruning is widely used for the reduced computational cost and the compatibility with off-the-shelf hardware devices. Among existing works, weights are typically removed using a predefined global threshold, or a threshold computed from a predefined metric. The predefined global threshold based designs ignore the variation among different layers and weights distribution, therefore, they may often result in sub-optimal performance caused by over-pruning or under-pruning. In this paper, we present a simple yet effective method, termed Optimal Thresholding (OT), to prune channels with layer dependent thresholds that optimally separate important from negligible channels. By using OT, most negligible or unimportant channels are pruned to achieve high sparsity while minimizing performance degradation. Since most important weights are preserved, the pruned model can be further fine-tuned and quickly converge with very few iterations. Our method demonstrates superior performance, especially when compared to the state-of-the-art designs at high levels of sparsity. On CIFAR-100, a pruned and fine-tuned DenseNet-121 by using OT achieves 75.99% accuracy with only 1.46e8 FLOPs and 0.71M parameters. code is available at: this https URL.
Submission history
From: Yun Ye [view email][v1] Tue, 10 Mar 2020 08:24:24 UTC (296 KB)
[v2] Thu, 12 Mar 2020 01:55:52 UTC (297 KB)
[v3] Fri, 22 May 2020 05:23:13 UTC (297 KB)
[v4] Sat, 27 Jun 2020 14:31:58 UTC (298 KB)
[v5] Thu, 10 Sep 2020 05:42:37 UTC (298 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.