SAS Viya Deep Learning - error when using batch normalization

[Helgard]

I am trying to fit a Deep Fully Connected Neural Network (DNN) in SAS Viya (3.4) and SAS Studio  4.4.  Just an example, I am trying to lear Viya and get up to speed.

Without batch normilization  all works well. However as soon as I add batch normalization layer:

AddLayer / model='BatchDLNN' name='HLayer2' layer={type='FULLCONNECT' n=30 act='identity' init='xavier' includeBias=False} srcLayers={'HLayer1'};
AddLayer / model='BatchDLNN' name='BatchLayer2' layer={type='BATCHNORM' act='ELU'} srcLayers={'HLayer2'};

I get the following error in the log:

ERROR: DNN models support only input, output, and fully connected layers.

ERROR: The action stopped due to errors.

Any suggestions?