This model was designed around Pycoco's dataset, the CNN model constructed outputs training loss graphs and a confusion matrix plotted using seaborn for the network of interest. This project was written for an assignment regarding computer vision in CNN. The project consists of 3 program files. The first being the main file, hw4_RussellHo which contains the script and imports from the two other programs; hw4_RussellHo inherits the network's structure from cnn_net1 containing the class for the CNN network and then is modified to include padding and 10 extra convolutional layers for net2 and net3 structures respectively. The third file, hw4_dataloader, contains the custom dataset-loader class written by myself in order to load the images into the dataloader by Pytorch.utils.data in batches.