Now I have developed an application written in Java. I want to call a model written and trained with python and gluon. The model has been trained and saved locally. How can I use Java to call this model in the developed application? ?Please help me.
Please see this thread, there are currently no java bindings.
However going forward you should be able to export your MXNet trained model to the ONNX format (not available yet, coming soon) and then run it using a different framework available in java, for example DeepLearning4J, though this is not currently possible.
MXNet now has java bindings for inference:
To read more about the Java inference API see the medium blogpost