helperImportMatConvNet error-- Supports only AlexNet

As I could observe, this SeriesNetwork object from the Neural Network Toolbox imports MatConvNet network into a SeriesNetwork object and the function supports only the AlexNet model from MatConvNet. Do we have this function restructured to support the other models in the Matconvnet toolbox?

 Akzeptierte Antwort

mizuki
mizuki am 29 Jan. 2017
Yes. You can for example load VGG16 with following code:
>> cnnURL = http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-f.mat;
>> cnnMatFile = 'imagenet-vgg-f.mat';
>> websave(cnnMatFile, cnnURL);
>> convnet = helperImportMatConvNet(cnnMatFile);

5 Kommentare

Thanks a lot for that info. Do we have this Series network object "helperImportMatConvNet" written for GoogleNet and Microsoft ResNet?
shivasakthi
shivasakthi am 2 Feb. 2017
Bearbeitet: Walter Roberson am 2 Feb. 2017
Hi, Against what you mentioned, it is not supporting the VGG 16 pre-trained model. it gives an error:
convnet = helperImportMatConvNet(cnnMatFile);
Error using helperImportMatConvNet>getPadding (line 233)
Only symmetric padding is supported
Error in helperImportMatConvNet>createPoolLayer (line 122)
padding = getPadding(l);
Error in helperImportMatConvNet (line 56)
layers{end+1} = createPoolLayer(l);
Kindly help with this.
dekwe
dekwe am 21 Jun. 2017
I have the same error when I want to use googleNet or resNet
I have the same error for % Specify folder for storing CNN model
cnnMatFile = 'C:\Users\Anand\Desktop\MATCONVNET\imagenet-vgg-f.mat';
% Load MatConvNet network into a SeriesNetwork
Net = helperImportMatConvNet(cnnMatFile);
as
"Error using helperImportMatConvNet>getPadding (line 233) Only symmetric padding is supported
Error in helperImportMatConvNet>createPoolLayer (line 122) padding = getPadding(l);
Error in helperImportMatConvNet (line 56) layers{end+1} = createPoolLayer(l);
Error in load_saved_network (line 22) Net = helperImportMatConvNet(cnnMatFile);"
Pls. help in this issue
Richa
Richa am 2 Apr. 2018
U can tweak this function of getPadding() to work for other models as well.

Melden Sie sich an, um zu kommentieren.

Weitere Antworten (0)

Kategorien

Mehr zu Deep Learning Toolbox finden Sie in Hilfe-Center und File Exchange

Gefragt:

am 27 Jan. 2017

Kommentiert:

am 2 Apr. 2018

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by