The previous chapter focused on training Multilayer Neural Networks (MNNs), and presenting code examples for CNNs and RNNs in particular. This chapter describes how monitoring a network can be done while training is in progress and how to use this monitoring information to tune a model. DL4J provides UI facilities for monitoring and tuning purposes, and will be the centerpiece of this chapter. These facilities also work in a training context with DL4J and Apache Spark. Examples for both situations (training using DL4J only and DL4J with Spark) will be presented. A list of potential baseline steps or tips for network training will also be discussed.
Monitoring and Debugging Neural Network Training
Get Hands-On Deep Learning with Apache Spark now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
