set labels for classifier deep learning toolbox

Where do i set the labels for supervised training?

[XTrain,YTrain] = digitTrain4DArrayData;

idx = randperm(size(XTrain,4),1000);
XValidation = XTrain(:,:,:,idx);
XTrain(:,:,:,idx) = [];
YValidation = YTrain(idx);
YTrain(idx) = [];

layers = [
    fullyConnectedLayer(4096,"Name","fc6","BiasInitializer","ones","WeightsInitializer","ones")
    reluLayer("Name","relu6")
    dropoutLayer(0.5,"Name","dropout6")
    fullyConnectedLayer(4096,"Name","fc7")
    reluLayer("Name","relu7")
    dropoutLayer(0.5,"Name","dropout7")
    fullyConnectedLayer(4096,"Name","fc8")
    softmaxLayer("Name","softmax")
    classificationLayer("Name","classoutput")];

plot(layerGraph(layers));

net = trainNetwork(XTrain,YTrain,layers,options);

Answer 1

[Ankriti Sachan]

By setting the labels for supervised training, I am assuming that you want to ask how to train the dataset with the labeled data.

I guess you are already following this documentation - https://www.mathworks.com/help/deeplearning/ref/trainnetwork.html

Check the section here - https://www.mathworks.com/help/deeplearning/ref/trainnetwork.html#d120e82918

For the function,

net = trainNetwork(X,Y,layers,options), it is mentioned that

X = Training data

Y = Labels of the data that you are training the model with

layers = Neural network layers

option = Training options.

So I guess ‘Y’ is the variable that you are looking for.

[Wiktor]

Hi! If I have augmentedImageDatastore created from imageDatadtore with defined Labels to every element, can I skip "Y" in trainNetwork function?