trainNetwork RESHAPE error line 170

Hi. I keep getting this error

Error using trainNetwork (line 170)

To RESHAPE the number of elements must not change.

I'm not quite sure what I am doing wrong since the trainNetwork function is kinda hard to understand without spending hours on going through it.

This is my code. My X is a 100x5 double and my Y is a 100x1 categorical.

layers = [sequenceInputLayer([100,5,1])

flattenLayer

bilstmLayer(100)

fullyConnectedLayer(10)

reluLayer

fullyConnectedLayer(2)

softmaxLayer

classificationLayer];

trained_net = train2(X,Y, layers);

Train2 function

function trained_net = train2(X,Y, layers)

% creates adam optimizer

options = trainingOptions('adam', 'InitialLearnRate',3e-4, ...

'SquaredGradientDecayFactor',0.99, 'MaxEpochs',20, ...

'MiniBatchSize',10, 'Plots','training-progress');

% trains network

trained_net = trainNetwork(X,Y,layers,options);

Answer 1

[Vinai Datta Thatiparthi]

Hello Elaine,

The code that you have written fails in the input validation phase of trainNetwork. The output size of the network does not match the size of the response data. Since you are defining a classification network, the output is a “classes mismatch” error. You may have to revisit the definition of layers to solve the problem.

Consider using the MATLAB debugging tools to pinpoint areas of code which is creating the error.

Hope this helps!

[Elaine Todd]

Hi Vinai,

Thanks for the answer. My code goes through iParseInputArguments fine, but the issue lies in the doTrainNetwork. I don't know how to change the output size of a classificationLayer as the definition for it does not say how to so I don't know how to fix the output size and response data not matching. I've narrowed down the error to line 260 in trainNetwork. I'm not quite sure how to get into trainer and see what is going wrong in it.

% Do the training

trainedNet = trainer.train(trainedNet, trainingDispatcher);

Thanks for the help!

[Vinai Datta Thatiparthi]

Hey Elaine,

Which version of MATLAB are you using?

[Namasi Gomathi]

Im facing the same problem on R2020a