Skip to content

Top-1 Acc=61.0% on ImageNet, without any sacrificing compared with SqueezeNet v1.1.

License

Notifications You must be signed in to change notification settings

miaow1988/SqueezeNet_v1.2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

SqueezeNet v1.2

Top-1 Acc=61.0% on ImageNet, without any sacrificing compared with SqueezeNet v1.1.

We improved SqueezeNet by the linear kernel compression idea, and got 61.0% top-1 accuracy on ImageNet.

In the original paper, each Fire model, which is the fundamental building block of SqueezeNet, consists of a squeeze part (a 1x1 convolutional layer with few filters) and a expend part (a 1x1 convolutional layer concatenated with a 3x3 convolutional layer, both of them have more filters compared with the squeeze part).

We removed the RELU after the squeeze layer in each Fire model. Now the Fire model can be seen as a new model consists of the expand part only, and both layers in the expand parts are linearly compressed by a same 1x1 convolutional layer.

In this way, we achieved 3% top-1 accuracy improvement on ImageNet based on SqueezeNet v1.1, without sacrificing parameter numbers and efficiency.

Implementation Details

The network was trained by MxNet. Image size is 224x224, and the pixel mean for preprocessing is mean_r=124, mean_g=117, mean_b=104. All other hyper parameters are as same as hyper parameters used in original SqueezeNet v1.1.

About

Top-1 Acc=61.0% on ImageNet, without any sacrificing compared with SqueezeNet v1.1.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published