We will be using neural network-based methods extensively in the following chapters. These methods are based on multidimensional array operations. TensorFlowTensorFlow provides a platform to perform all these operations. The KerasKeras API will simplify life for use while benefiting from TensorFlow.TensorFlow Therefore, we will consider both in this chapter. To do so, we will start with their installation. Then, we will explore constant and variable declarations in TensorFlow. Afterward, we will define arithmetic operations on these. Then, we will consider activation functions under TensorFlow. We will also provide the link between the NumPy library and TensorFlow. We will consider loading data as the first end of chapter application. Next, we will introduce loading and storing neural network models under TensorFlow as the second end of chapter application. Finally, we will evaluate converting neural network models from other platforms to TensorFlow format as the third end of chapter application.