In this chapter, we have learned how to programmatically evaluate a model's efficiency using the different facilities that are provided by the DL4J API. We have now closed the full circle in terms of the implementation, training, and evaluation of MNN using DL4J and Apache Spark.
The next chapter will give us some insight into the deployment of a distribution environment and importing and executing pre-trained Python models, as well as a comparison of DL4J with some alternative DL frameworks for the Scala programming language.
