The crucial feature of C++ is the ability of the program to be compiled and run on a variety of hardware platforms. You can train your complex machine learning model on the fastest GPU in the data center, and deploy it to tiny mobile devices with limited resources. This chapter will show you how to use the C++ APIs of the TensorFlow and Caffe2 libraries to build programs that use machine learning models for Android devices and in the cloud.

This section comprises the following chapters:

• Chapter 12, Exporting and Importing Models
• Chapter 13, Deploying Models on Mobile and Cloud Platforms