Matlab Cross-Validation error in calculating performance of model

I am trying to perform cross-validation on my models with multiple classification classes but I am getting errors when trying to update my classperf on each fold. I am getting this error "Index vector has invalid values."

My code is as below:

K = 5; 
N = size(DataSet, 1); 
idx = crossvalind('Kfold', N, K); 
cp = classperf(trainLabel);
for i = 1:K
   ... 
   %Long codes for svmtrain & svmclassify
   ...
   cp = classperf(cp, Group, trueTestLabel); %error on this line
end
cp.CorrectRate

where trainLabel is a 120 x 1 double consisting of the ground-truth score for all of the items; Group is the results obtained from svmclassify in 20 x 1 double; and trueTestLabel is the ground-truth score for the testing class in 20 x 1 double form obtained using the following function:

trueTestLabel = trainLabel(idx == i, end); I have tried converting "Group" and "trueTestLabel" into cells by using num2cell function as such:

cp = classperf(cp, num2cell(Group), num2cell(trueTestLabel ));

but instead, I got a different error saying "When the class labels of the CP object are numeric, the output of the classifier must be all non-negative integers or NaN's."