categorization <undefined>

Hi

I have the following MATLAB script - see below

When I run the script I get all categorization as <undefinded>

Unclear why - there is a reasonable coverage of all six categories during the training session

I have attached the input I am using

Cheers

Bob M

%% Implement and test the neural net to classify synthetic control data into six categories
%% Inputs
% s6 - historical data - 600*1 cell array
%         each cell comprising - 60*1 double              
% t6 - categories - 600*1 double [6 of - normal, cyclic, increasing trend, 
%                         decreasing trend, upward shift & downward shift] 

%% Split the data into training and testing data sets
n = length(s6);
nTrain = floor(0.75*n);
XTrain = s6(1:nTrain);
YTrain = categorical(t6(1:nTrain));
XTest = s6(nTrain+1:n);
YTest = categorical(t6(nTrain+1:n));

%% Visualize an example of each of the six categories.
figure
plot(XTrain{1}')
hold on
plot(XTrain{2}')
hold on
plot(XTrain{18}')
hold on
plot(XTrain{13}')
hold on
plot(XTrain{3}')
hold on
plot(XTrain{11}')
xlabel("Time Steps")
title("Six Training Observations")
legend('normal', 'cyclic', 'incr. trend', 'decr. trend', 'upward shift', 'downward shift', 'Location','northeastoutside')

%% Define the layers for the net
% This gives the structure of the convolutional neural net
numFeatures = 60;
NumHiddenUnits = 10;
numClasses = 6;
layers = [ ...
    sequenceInputLayer(numFeatures)
    lstmLayer(numHiddenUnits,'OutputMode','last')
    tanhLayer
    fullyConnectedLayer(numClasses)
    softmaxLayer
    classificationLayer]
maxEpochs = 10;
options = trainingOptions('adam', ...
    'InitialLearnRate', 0.001, ...
    'ExecutionEnvironment','cpu', ...
    'GradientThresholdMethod', 'l2norm', ...
    'GradientThreshold',0.5, ...
    'MaxEpochs',maxEpochs, ...
    'Verbose',0, ...
    'Plots','training-progress');

%% Analyse the network
analyzeNetwork(layers)
disp(layers)

%% Training   
%options = trainingOptions('adam', ...
%    'MaxEpochs',40, ...
%    'GradientThreshold', 1, ...
%    'Verbose', 0, ...
%    'Plots','training-progress');
net = trainNetwork(XTrain,YTrain,layers,options);

%% Test the neural net
YPred  = classify(net,XTest);

%% display the predictions against the test categories
disp(YTest);
disp(YPred);
accuracy = sum(YPred == YTest)/numel(YTest);
fprintf('Accuracy is %8.2f%%\n',accuracy*100)