CNN-Classification

This repository contains CNN classification code for the MNIST and Fashion MNIST datasets, implemented using the PyTorch library in Python. Approximately 99.4% classification accuracy is achieved for the MNIST dataset, while around 92.3% classification accuracy is achieved for the Fashion MNIST dataset. The network used for the Fashion MNIST dataset is deeper compared to the one used for the MNIST dataset, and therefore it is trained for more epochs and takes more training time. GPU is utilized for speed improvement.