我正在使用CNN中的Resnet50训练数据,但是数据过拟合。我想减少过度拟合。所以我想添加正则化L2。有人可以告诉我如何在代码中添加L2吗?您可以在下面看到我的代码。
clear all
close all
imds = imageDatastore("E:\test\data", ...
'IncludeSubfolders',true,'LabelSource','foldernames');
[imdsTrain,imdsValidation] = splitEachLabel(imds,0.7,'randomize'); %70% for train 30% for test
net=resnet50; % for the first time,you have to download the package from Add-on explorer
%Replace Final Layers
numClasses = numel(categories(imdsTrain.Labels));
lgraph = layerGraph(net);
newFCLayer = fullyConnectedLayer(numClasses,'Name','new_fc','WeightLearnRateFactor',10,'BiasLearnRateFactor',10);
lgraph = replaceLayer(lgraph,'fc1000' ,newFCLayer);
newClassLayer = classificationLayer('Name','new_classoutput');
lgraph = replaceLayer(lgraph,'ClassificationLayer_predictions',newClassLayer);
%Train Network
inputSize = net.Layers(1).InputSize;
augimdsTrain = augmentedImageDatastore(inputSize(1:2),imdsTrain);
augimdsValidation = augmentedImageDatastore(inputSize(1:2),imdsValidation);
options = trainingOptions('sgdm', ...
'MiniBatchSize',10, ...
'MaxEpochs',20, ...
'InitialLearnRate',1e-3, ...
'Shuffle','every-epoch', ...
'ValidationData',augimdsValidation, ...
'ValidationFrequency',5, ...
'Verbose',false, ...
'Plots','training-progress');
trainedNet = trainNetwork(augimdsTrain,lgraph,options);
YPred = classify(trainedNet,augimdsValidation);
accuracy = mean(YPred == imdsValidation.Labels)
C = confusionmat(imdsValidation.Labels,YPred)
cm = confusionchart(imdsValidation.Labels,YPred);
cm.Title = 'Confusion Matrix for Validation Data';
cm.ColumnSummary = 'column-normalized';
cm.RowSummary = 'row-normalized';
答案 0 :(得分:0)
在ResNet50 CNN模型上实施L2正则化:
resnet_base = ResNet50(weights='imagenet', include_top=False, input_shape=(224,224,3))
alpha = 0.00002
for layer in resnet_base.layers:
if isinstance(layer, keras.layers.Conv2D) or isinstance(layer, keras.layers.Dense):
layer.add_loss(keras.regularizers.l2(alpha)(layer.kernel))
if hasattr(layer, 'bias_regularizer') and layer.use_bias:
layer.add_loss(keras.regularizers.l2(alpha)(layer.bias))
希望这会有所帮助!