Timezone: »

 
Poster
Structural Pruning via Latency-Saliency Knapsack
Maying Shen · Hongxu Yin · Pavlo Molchanov · Lei Mao · Jianna Liu · Jose M. Alvarez

@
Structural pruning can simplify network architecture and improve inference speed. We propose Hardware-Aware Latency Pruning (HALP) that formulates structural pruning as a global resource allocation optimization problem, aiming at maximizing the accuracy while constraining latency under a predefined budget on targeting device. For filter importance ranking, HALP leverages latency lookup table to track latency reduction potential and global saliency score to gauge accuracy drop. Both metrics can be evaluated very efficiently during pruning, allowing us to reformulate global structural pruning under a reward maximization problem given target constraint. This makes the problem solvable via our augmented knapsack solver, enabling HALP to surpass prior work in pruning efficacy and accuracy-efficiency trade-off. We examine HALP on both classification and detection tasks, over varying networks, on ImageNet and VOC datasets, on different platforms. In particular, for ResNet-50/-101 pruning on ImageNet, HALP improves network throughput by $1.60\times$/$1.90\times$ with $+0.3\%$/$-0.2\%$ top-1 accuracy changes, respectively. For SSD pruning on VOC, HALP improves throughput by $1.94\times$ with only a $0.56$ mAP drop. HALP consistently outperforms prior art, sometimes by large margins. Project page at \url{https://halp-neurips.github.io/}.
Keywords: model compression · deep neural network pruning · latency reduction
[ Paper]  [ Poster]  [ OpenReview] 

Author Information

Maying Shen (NVIDIA)
Hongxu Yin (NVIDIA)
Pavlo Molchanov (NVIDIA)

Research scientist at NVIDIA from May 2015. Received PhD in radar target classification from Tampere University of Technology, Finland, in 2014.

Lei Mao (NVIDIA)
Jianna Liu
Jose M. Alvarez (NVIDIA)

More from the Same Authors