Setup of a distributed environment with DeepLearning4j

This section explains some tricks to do when setting up a production environment for DL4J neural network model training and execution.

Memory management

In Chapter 7, Training Neural Networks with Spark, in the Performance considerations section, we learned how DL4J handles memory when training or running a model. Because it relies on ND4J, it also utilizes off-heap memory and not only heap memory. Being off-heap, it means that it is outside the scope managed by the JVM's Garbage Collection (GC) mechanism (the memory is allocated outside the JVM). At the JVM level, there are only pointers to off-heap memory locations; they can be passed to the C++ code via the Java...