Java Deep Learning Cookbook

How it works...

Although we used save() or load() for the model's persistence in a local machine, it is not an ideal practice in production. For a distributed cluster environment, we can make use of BufferedInputStream/BufferedOutputStream in steps 1 and 2 to save/load models to/from clusters. We can use ModelSerializer or save()/load() just like we demonstrated earlier. We just need to be aware of the cluster resource manager and model persistence, which can be performed across clusters.