ERA V2 Assignment 5 was related to MNIST Dataset
The MNIST database is a large database of handwritten digits
We created a basic model to train it
In utils.py file you will find helper functions like defining train and test transformations, train and tes functions and training and test loss and accuracy
In models.py file you will find architecture of CNN, in the output layer we have used softmax because the output labels are 10. So, when the output lables are more then 2 then its a multi-class problem so at that time we generally used softmax activation function
We have then imported this utils and models files in our S5 ipython notebook to run our whole code