AttributeTypeException, Keras changes type of label to rm_metadata(

naveen_paul

I have the keras extension executing normally with rm_metadata("real", prediction") on a few runs for a numerical attribute and then tries to predict a nominal value for the same attribute for the same dataset and gives an error on other runs. 

hughesfleming68

Are you just changing the label and not making any other changes to the operators inside the Keras operator? I would post your process. Alternatively, you could also try your process with the new Deep Learning 4J extension.

My guess is that you would have to make changes to your output layer inside the Keras network. I have had some issues with the Keras extension. I have gone back to using Tensorflow just with Python.