Skip to content

NidabaSys/Simplified_SqueezeNet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 

Repository files navigation

Simplified_SqueezeNet

An improved version of SqueezeNet networks https://github.com/DeepScale/SqueezeNet

Simplified SqueezeNet is SqueezeNet network with removing Fire/Expand layer of reception field 1 and the concatenation layer. The suggested modification accelerated the training of SqueezeNet by more than 4 times.

To understand the differences between the two networks, refer the the below links.

This link visualize the SqueezeNet network: http://ethereon.github.io/netscope/#/gist/d2bf1b838c4c07a465afbdcfac0577f9

This link visualize the Simplified SqueezeNet network: http://ethereon.github.io/netscope/#/gist/fd7f749c3cfe69a6859cc56225c10c0f

The above directories contain the source files for training SqueezeNet and Simplified SqueezeNet on two large scale image datasets.

Stay tuned for the technical report (coming soon)

For any information , please contact us at [email protected]

If you find the network useful in your research, please consider citing:

@article{SqueezeNet,
    Author = {Hussein Al-barazanchi},
    Title = {Simplified SqueezeNet},
    Journal = {arXiv},
    Year = {2017}
}