How can I implement SVM as classifier on deeplearning stack?

Hello MATLAB community,

I have read few papers which use linear SVM as classifier on deep learning network. For ex.

https://arxiv.org/pdf/1306.0239.pdf

Can someone please suggest me how can I also replace cross-entropy (or softmax) and implement SVM as classification layer over deep learning network.

I have been suggested that we can define a custom loss layer. An example is available in MATLAB documentation here

https://www.mathworks.com/help/deeplearning/ug/define-custom-classification-output-layer.html

Thanks in advance

[Haijun Su]

I have the same question.

[SG]

Did you find the answer yet.