We already see the Keras API in Tensorflow 2.x samples, that is the Tensorflow built-in Keras utilities. Here we try to use the standard-alone Keras Library Framework to illustrate the usage of deep learning architecture. Keras library provides the front-end APIs, the back-end library can be Theano, Tensorflow, etc. In these sample, we use Tensorflow as the back-end support library. See the sample for mode details such as Sequential model definition, optimizer selection, model compilation, and training and evaluation.