MATLAB Answers

L1 and L2 Regularization for matlab

57 views (last 30 days)
Abdussalam Elhanashi
Abdussalam Elhanashi on 17 Oct 2019
Answered: Divya Gaddipati on 21 Oct 2019
Hi Guys
I would like to know how to add regularization L1 & L2 for following layers to reduce overfitting
imageInputLayer([32 32 3],"Name","imageinput")
convolution2dLayer([5 5],32,"Name","conv","BiasLearnRateFactor",2,"Padding",[2 2 2 2],"WeightsInitializer","narrow-normal")
batchNormalizationLayer
maxPooling2dLayer([3 3],"Name","maxpool","Stride",[2 2])
preluLayer(20,'prelu')
averagePooling2dLayer([3 3],"Name","avgpool","Stride",[2 2])
fullyConnectedLayer(2,"Name","fc_rcnn","BiasL2Factor",1,"BiasLearnRateFactor",10,"WeightLearnRateFactor",20,"WeightsInitializer","narrow-normal")
dropoutLayer(0.65,'Name','drop1')
softmaxLayer("Name","softmax")
classificationLayer("Name","classoutput")];

  0 Comments

Sign in to comment.

Answers (1)

Divya Gaddipati
Divya Gaddipati on 21 Oct 2019
You can set the L2 regularization for selected layers using the setl2factor function.
You can refer to the following link for more understanding:

  0 Comments

Sign in to comment.

Sign in to answer this question.