Devised pipelines for the task of Digit Classification using Deep Learning on MNIST dataset.

The repository contains 2 codes:

1. MNIST(NumPy):

• Built a simple Neural Network from scratch using NumPy library.
• Achieved 97.07% accuracy.
• Plot for loss vs no. of iterations is as follows

1. MNIST(PyTorch):

• Built a Convolution Neural Network(CNN) for the same using PyTorch library.
• Achieved 98.87% accuracy.
• Plot for loss vs no. of iterations is as follows