Practical Automated Machine Learning Using H2O.ai

11

Working with Model Object, Optimized (MOJO)

As we learned in Chapter 10, Working with Plain Old Java Objects (POJOs), when working with production systems, we need portable software that we can easily deploy to our production servers. It is especially important in Machine Learning (ML) services that ML models be portable and self-sufficient. This helps engineers deploy new models regularly without worrying about breaking their systems in production because of any dependency issues.

H2O’s model POJOs were a good solution to this problem. Model POJOs are H2O models that can be extracted in the form of Java POJOs that you can directly run using Java Virtual Machine (JVM) with the help of h2o-genmodel.jar.

However, model POJOs have certain drawbacks ...