A custom trained CNN will be implemented for the task of classification using the CIFAR-10 dataset which consists of 10,000 images of 28x28 pixels. The CNN will then be tuned and optimized and compared against other leading edge competitors such as ResNet-18, MobileNet and AlexNet which are all modified to suit the input image size. What this experiment and research has found is that a custom trained CNN can outperform the other competitors as long as it is tuned appropriately and each models have their strengths and weaknesses, and will suit their own appropriate use cases.