Keras – 9.0 Introduction to Network Architecture

Welcome to CS With James

In this introduction I will discuss about what kind of network architectures that I will discuss in the next few day.

Before starting the introduction, all the network architectural design is based on the Image-Net classification competition. Image-net takes 224x224x3 as an input size, so we cannot use exactly the same network from the paper. I will modify the filter size and the strides to make the network works with the CIFAR10.

These are the list that I will talk about in this series.

  • AlexNet
  • VGGNet
  • GoogLeNet (Inception)
  • DenseNet

There are lot more famous network architectures but these four architecture is the most famous and most important ones.

I will code each of the Networks and train 1k epochs on the CIFAR10 dataset and compare each other. Of course I will explain how it works and why it is important in network architecture.

I will explain the details in the each post but here is the visualization of each network architectures.

  1. AlexNet
  2. VGGNet
  3. GoogLeNet
  4. DenseNet


Leave a Reply

Your email address will not be published. Required fields are marked *