Neural Networks using PyTorch

1. A simple multilayer Perceptron (MLP) or simply one hidden layer neural network (NN) developed using PyTorch. It is trained and tested on MNIST dataset with average test accuracy around 95%.

2. A 2D Convolutional Neural Net (CNN) developed using pytorch. It is trained and tested on MNIST dataset. It reaches accuracy level over 99% after 10 epochs. The train loss, test loss and test accuracy with epoch number have been plotted, in Jupyter Notebook.