Image classifier using CNN in PyTorch Lightning and MNIST data

The CNN_MNIST_PyTorch_Lightning.ipynb notebook shows how to train a CNN (including testing it and using it to make predictions) to predict the label on the MNIST database which is a set of images with handwritten digts from 0 to 9. The trained model reaches ~99% accuracy with only 6 epochs.

Dependencies

torch
torchmetrics
torchvision
pytorch-lightning
sklearn
matplotlib