Simple NN

Simple NN is exactly what it says. It is a simplistic implementation of a dense neural network layer.

To demonstrate the usage of the implementation, it is used to train a dense neural network to recognize handwritten digits using the MNIST dataset.

The trained model achieves 99.3% test accuracy on Kaggle.

Required dependencies

  • python3
  • numpy
  • pandas
  • matplotlib
  • jupyter notebook

Running the program

  • Open Jupyter notebook. It will open in your browser.
  • Browse to the location of the project.
  • Open the file MNIST.ipynb
  • Run each cell by pressing Shift + Enter. It will show the output of the cell, if any.