Error in weighted classification layer implementation (following matlab documentation)

I am following Matlab documentation to implement a custom weighted cross-entropy loss function:

https://nl.mathworks.com/help/deeplearning/ug/create-custom-weighted-cross-entropy-classification-layer.html#mw_f8874094-d1f5-4356-839d-bfeaa1bb710e

However, when I try the code on R2018b I get the following error:

No appropriate method, property, or field 'OutputSize' for class 'weightedClassificationLayer'

Anybody knows what the problem is?

My layer sequence:

layers = [ ...
    sequenceInputLayer(dimPattern)
    lstmLayer(numHiddenUnits,'OutputMode','last')
    fullyConnectedLayer(numClasses)
    softmaxLayer
    weightedClassificationLayer(lossWeights) 
    ];

Maybe handy: I am using a mini-batch datastore as input to "trainNetwork", does it make a difference?