in the previous lesson we saw how we can build and train a deep neural network to classify images of articles of clothing using the fashion and this dataset and the results we achieved were pretty impressive considering we only needed a few lines of code to be able to do so you also experimented with the impact that the number of units in each layer and the number of training epochs had on the final training accuracy but how do you make this model perform even better one way is to use something called a convolutional neural network which
are commonly abbreviated as CNN's CNN's have proven to achieve higher accuracies when classifying images than the simple pure dense neural networks we saw in the previous lessons for this reason CNN's have become extremely useful and they're actually the primary reason for breakthroughs in computer vision applications using machine learning in this lesson we'll see how easy it is to create a CNN based image classifier from scratch using tensorflow and caris to make things familiar we will reuse the fashion Abney's dataset and create a model that uses a CNN we can then compare how well our CNN
performs compared to the neural network we created in the previous lesson but before we jump into the code let's first take some time to understand how convolutional neural networks work the two main concepts in a convolutional neural network are convolutions armax bullying so let's start by taking a look at compositions
