Recent research on deep neural networks has focused primarily on improving
accuracy. For a given accuracy level, it is typically possible to identify
multiple DNN architectures that achieve that accuracy level. With equivalent
accuracy, smaller DNN architectures offer at least three advantages: (1)
Smaller DNNs require less communication across servers during distributed
training. (2) Smaller DNNs require less bandwidth to export a new model from
the cloud to an autonomous car. (3) Smaller DNNs are more feasible to deploy on
FPGAs and other hardware with limited memory. To provide all of these
advantages, we propose a small DNN architecture called SqueezeNet. SqueezeNet
achieves AlexNet-level accuracy on ImageNet with 50x fewer parameters.
Additionally, with model compression techniques we are able to compress
SqueezeNet to less than 0.5MB (510x smaller than AlexNet).
The SqueezeNet architecture is available for download here:
https://github.com/DeepScale/SqueezeNet
%0 Generic
%1 iandola2016squeezenet
%A Iandola, Forrest N.
%A Han, Song
%A Moskewicz, Matthew W.
%A Ashraf, Khalid
%A Dally, William J.
%A Keutzer, Kurt
%D 2016
%K cnn compression deep_learning model_reduction topology sparse sparsity
%T SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB
model size
%U http://arxiv.org/abs/1602.07360
%X Recent research on deep neural networks has focused primarily on improving
accuracy. For a given accuracy level, it is typically possible to identify
multiple DNN architectures that achieve that accuracy level. With equivalent
accuracy, smaller DNN architectures offer at least three advantages: (1)
Smaller DNNs require less communication across servers during distributed
training. (2) Smaller DNNs require less bandwidth to export a new model from
the cloud to an autonomous car. (3) Smaller DNNs are more feasible to deploy on
FPGAs and other hardware with limited memory. To provide all of these
advantages, we propose a small DNN architecture called SqueezeNet. SqueezeNet
achieves AlexNet-level accuracy on ImageNet with 50x fewer parameters.
Additionally, with model compression techniques we are able to compress
SqueezeNet to less than 0.5MB (510x smaller than AlexNet).
The SqueezeNet architecture is available for download here:
https://github.com/DeepScale/SqueezeNet
@misc{iandola2016squeezenet,
abstract = {Recent research on deep neural networks has focused primarily on improving
accuracy. For a given accuracy level, it is typically possible to identify
multiple DNN architectures that achieve that accuracy level. With equivalent
accuracy, smaller DNN architectures offer at least three advantages: (1)
Smaller DNNs require less communication across servers during distributed
training. (2) Smaller DNNs require less bandwidth to export a new model from
the cloud to an autonomous car. (3) Smaller DNNs are more feasible to deploy on
FPGAs and other hardware with limited memory. To provide all of these
advantages, we propose a small DNN architecture called SqueezeNet. SqueezeNet
achieves AlexNet-level accuracy on ImageNet with 50x fewer parameters.
Additionally, with model compression techniques we are able to compress
SqueezeNet to less than 0.5MB (510x smaller than AlexNet).
The SqueezeNet architecture is available for download here:
https://github.com/DeepScale/SqueezeNet},
added-at = {2018-06-16T08:51:27.000+0200},
author = {Iandola, Forrest N. and Han, Song and Moskewicz, Matthew W. and Ashraf, Khalid and Dally, William J. and Keutzer, Kurt},
biburl = {https://www.bibsonomy.org/bibtex/2f68f9d56e6b2db8cef4292414eafbb87/loroch},
description = {1602.07360.pdf},
interhash = {d407ce69b22e9fc2721bed228dad9f68},
intrahash = {f68f9d56e6b2db8cef4292414eafbb87},
keywords = {cnn compression deep_learning model_reduction topology sparse sparsity},
note = {cite arxiv:1602.07360Comment: In ICLR Format},
timestamp = {2018-06-16T10:46:03.000+0200},
title = {SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB
model size},
url = {http://arxiv.org/abs/1602.07360},
year = 2016
}