Clear Filters
Clear Filters

unrecognized method property or field Labels for class augmentdatastore?

19 views (last 30 days)
I am tring to train the model on .mat dataset. i have train the model sucessfully but when i tried to find the accuracy i got the error.
imds = imageDatastore('D:\yellow\img-data\iqmat\', 'FileExtensions', '.mat', 'IncludeSubfolders',true, ...
'LabelSource','foldernames',...
'ReadFcn',@matReader);
[imdsTrain,imdsValidation] = splitEachLabel(imds,0.7, 'randomized');
inputSize = lgraph_1.Layers(1).InputSize;
[learnableLayer,classLayer] = findLayersToReplace(lgraph_1);
numClasses = numel(categories(imdsTrain.Labels));
if isa(learnableLayer,'nnet.cnn.layer.FullyConnectedLayer')
newLearnableLayer = fullyConnectedLayer(numClasses, ...
'Name','new_fc', ...
'WeightLearnRateFactor',10, ...
'BiasLearnRateFactor',10);
elseif isa(learnableLayer,'nnet.cnn.layer.Convolution2DLayer')
newLearnableLayer = convolution2dLayer(1,numClasses, ...
'Name','new_conv', ...
'WeightLearnRateFactor',10, ...
'BiasLearnRateFactor',10);
end
lgraph_1 = replaceLayer(lgraph_1,learnableLayer.Name,newLearnableLayer);
newClassLayer = classificationLayer('Name','new_classoutput');
lgraph_1 = replaceLayer(lgraph_1,classLayer.Name,newClassLayer);
imdsTrain = augmentedImageDatastore([224,224],imdsTrain);
imdsValidation = augmentedImageDatastore([224,224],imdsValidation);
miniBatchSize =8;
valFrequency = floor(numel(imdsTrain.Files)/miniBatchSize);
checkpointPath = pwd;
options = trainingOptions('sgdm', ...
'MiniBatchSize',miniBatchSize, ...
'MaxEpochs',100, ...
'InitialLearnRate',1e-4, ...
'Shuffle','every-epoch', ...
'ValidationData',imdsValidation, ...
'ValidationFrequency',valFrequency, ...
'Verbose',false, ...
'Plots','training-progress', ...
'CheckpointPath',checkpointPath,...
'ExecutionEnvironment','gpu');
net = trainNetwork(imdsTrain,lgraph_1,options);
[YPred,probs] = classify(net,imdsValidation);
accuracy = mean(YPred == imdsValidation.Labels)
error:
unrecognized method property or field Labels for class augmentdatastore

Accepted Answer

Walter Roberson
Walter Roberson on 14 Dec 2021
augmentedImageDatastore() does not record the labels of the input data store.
You currently have
imdsValidation = augmentedImageDatastore([224,224],imdsValidation);
which takes imdsValidation (an image data store that has labels) as input, and you write to the same variable... but augmentedImageDatastore does not carry the labels.
If you wrote to a different variable, then when you got to
accuracy = mean(YPred == imdsValidation.Labels)
you could be referring to the unaugmented data store that still has the labels.
  6 Comments
Walter Roberson
Walter Roberson on 15 Dec 2021
imds = imageDatastore('D:\yellow\img-data\iqmat\', 'FileExtensions', '.mat', 'IncludeSubfolders',true, ...
'LabelSource','foldernames',...
'ReadFcn',@matReader);
[imdsTrain,imdsValidation] = splitEachLabel(imds,0.7, 'randomized');
inputSize = lgraph_1.Layers(1).InputSize;
[learnableLayer,classLayer] = findLayersToReplace(lgraph_1);
numClasses = numel(categories(imdsTrain.Labels));
if isa(learnableLayer,'nnet.cnn.layer.FullyConnectedLayer')
newLearnableLayer = fullyConnectedLayer(numClasses, ...
'Name','new_fc', ...
'WeightLearnRateFactor',10, ...
'BiasLearnRateFactor',10);
elseif isa(learnableLayer,'nnet.cnn.layer.Convolution2DLayer')
newLearnableLayer = convolution2dLayer(1,numClasses, ...
'Name','new_conv', ...
'WeightLearnRateFactor',10, ...
'BiasLearnRateFactor',10);
end
lgraph_1 = replaceLayer(lgraph_1,learnableLayer.Name,newLearnableLayer);
newClassLayer = classificationLayer('Name','new_classoutput');
lgraph_1 = replaceLayer(lgraph_1,classLayer.Name,newClassLayer);
imdsTrain = augmentedImageDatastore([224,224],imdsTrain);
imdsValidation_aug = augmentedImageDatastore([224,224],imdsValidation); %HERE
miniBatchSize =8;
valFrequency = floor(numel(imdsTrain.Files)/miniBatchSize);
checkpointPath = pwd;
options = trainingOptions('sgdm', ...
'MiniBatchSize',miniBatchSize, ...
'MaxEpochs',100, ...
'InitialLearnRate',1e-4, ...
'Shuffle','every-epoch', ...
'ValidationData',imdsValidation_aug, ... %HERE
'ValidationFrequency',valFrequency, ...
'Verbose',false, ...
'Plots','training-progress', ...
'CheckpointPath',checkpointPath,...
'ExecutionEnvironment','gpu');
net = trainNetwork(imdsTrain,lgraph_1,options);
[YPred,probs] = classify(net,imdsValidation_aug);
accuracy = mean(YPred == imdsValidation.Labels)

Sign in to comment.

More Answers (0)

Community Treasure Hunt

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

Start Hunting!