English

SqueezeNext: Hardware-Aware Neural Network Design

Neural and Evolutionary Computing 2021-04-21 v2

Abstract

One of the main barriers for deploying neural networks on embedded systems has been large memory and power consumption of existing neural networks. In this work, we introduce SqueezeNext, a new family of neural network architectures whose design was guided by considering previous architectures such as SqueezeNet, as well as by simulation results on a neural network accelerator. This new network is able to match AlexNet's accuracy on the ImageNet benchmark with 112×112\times fewer parameters, and one of its deeper variants is able to achieve VGG-19 accuracy with only 4.4 Million parameters, (31×31\times smaller than VGG-19). SqueezeNext also achieves better top-5 classification accuracy with 1.3×1.3\times fewer parameters as compared to MobileNet, but avoids using depthwise-separable convolutions that are inefficient on some mobile processor platforms. This wide range of accuracy gives the user the ability to make speed-accuracy tradeoffs, depending on the available resources on the target hardware. Using hardware simulation results for power and inference speed on an embedded system has guided us to design variations of the baseline model that are 2.59×2.59\times/8.26×8.26\times faster and 2.25×2.25\times/7.5×7.5\times more energy efficient as compared to SqueezeNet/AlexNet without any accuracy degradation.

Keywords

Cite

@article{arxiv.1803.10615,
  title  = {SqueezeNext: Hardware-Aware Neural Network Design},
  author = {Amir Gholami and Kiseok Kwon and Bichen Wu and Zizheng Tai and Xiangyu Yue and Peter Jin and Sicheng Zhao and Kurt Keutzer},
  journal= {arXiv preprint arXiv:1803.10615},
  year   = {2021}
}

Comments

12 Pages

R2 v1 2026-06-23T01:07:45.516Z