We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration

Abstract: Previous works utilized ''smaller-norm-less-important'' criterion to prune filters with smaller norm values in a convolutional neural network. In this paper, we analyze this norm-based criterion and point out that its effectiveness depends on two requirements that are not always met: (1) the norm deviation of the filters should be large; (2) the minimum norm of the filters should be small. To solve this problem, we propose a novel filter pruning method, namely Filter Pruning via Geometric Median (FPGM), to compress the model regardless of those two requirements. Unlike previous methods, FPGM compresses CNN models by pruning filters with redundancy, rather than those with ''relatively less'' importance. When applied to two image classification benchmarks, our method validates its usefulness and strengths. Notably, on CIFAR-10, FPGM reduces more than 52% FLOPs on ResNet-110 with even 2.69% relative accuracy improvement. Moreover, on ILSVRC-2012, FPGM reduces more than 42% FLOPs on ResNet-101 without top-5 accuracy drop, which has advanced the state-of-the-art. Code is publicly available on GitHub: this https URL
Comments: Accepted to CVPR 2019 (Oral)
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:1811.00250 [cs.CV]
  (or arXiv:1811.00250v3 [cs.CV] for this version)

Submission history

From: Yang He [view email]
[v1] Thu, 1 Nov 2018 06:03:05 GMT (253kb,D)
[v2] Sat, 6 Apr 2019 03:59:11 GMT (794kb,D)
[v3] Sun, 14 Jul 2019 10:31:03 GMT (809kb,D)

Link back to: arXiv, form interface, contact.