Handwritten digit recognition using neural network, trained on 60000 images from MNIST dataset.
The model achieved an accuracy over 97% tested on 10000 images. It's built from scratch using PyTorch.
Read the step-by-step detailed article here on Towards Data Science.
For queries, feel free to drop a message on LinkedIn.

