Sounds like there would be a difference using the model for inference, because Gluon version will not have a softmax as the output of the model.
I recommend providing an inference example for both gluon and symbol: VI) Inference using the model [ Full content available at: https://github.com/apache/incubator-mxnet/pull/12542 ] This message was relayed via gitbox.apache.org for [email protected]
