/mnist-classifier-web-app

Primary LanguagePython

mnist-classifier-web-app

This is one of my early projects, finished by the end of dec 2022

This is a web app that can classify handwritten digits from 0 to 9. The model is trained using the MNIST dataset. The model is a feed forward neural network with 2 hidden layers. The model is trained using PyTorch. The web app is built using Flask.

Dependencies

  • flask
  • torch
  • torchvision
  • matplotlib

How To Use

  1. In order to get the .pth file, run python app/mnist_ffn.py to train the model
  2. Run python app/app.py to start the web app
  3. Run python app/test_six.py to test the web app

References