In this chapter, we learned how to create convolutional neural networks with TensorFlow and Keras. We learned the core concepts of convolution and pooling, that lay the foundation of CNN. We learned the LeNet family of architectures and created, trained, and evaluated the LeNet family model for MNIST and CIFAR datasets. TensorFlow and Keras offer many convolutional and pooling layers and operations. The reader is encouraged to explore the layers and operations that were not covered in this chapter.
In the next chapter, we shall continue our journey to learn how to apply TensorFlow on image data with the AutoEncoder architecture.
