This model uses Convolutional Neural Networks to classify handwritten digits from the keras MNIST dataset.
I could achieve an accuracy of 0.99 on both training and test set.
Use of GPU is recommended for this project as it cuts the execution time by a factor of 10