AI Sudoku

AI Sudoku solver using CNN/KNN

This project uses convolutional neural networks to analyze a sudoku image and converts it to a digital format

Index

• Getting Started
• Running the App

Getting Started

1. Download and install Python3 from this link
2. Install venv to create a virtual environment for the project.
   • You can do this using a terminal and type :

     py -m pip install --user virtualenv

   • For macOS and Linux:

     python3 -m pip install --user virtualenv

3. Now create and activate the virtual environment

   venv Sudoku

   • For Windows

     cd Sudoku\Scripts
     activate

   • For macOS and Linux

     source Sudoku/bin/activate

4. Clone the repo, install the required modules using requirements.txt

   pip3 install -r requirements.txt

   • For Apple Silicon/ M-series Mac users, please check out Apple's Tensorflow Metal installation guide

Running the App

• Before running the App in itself, you can switch the model type between CNN and KNN.

  from UI import UI
  from ConvNet import CNN
  from KNN import KNN
  import os
  # Change the model type variable value to "CNN" to use the Convolutional Neural Network
  # Change the model type variable value to "KNN" to use the K Nearest Neighbours Classifier
  modeltype = "CNN"

• Head Over and run Start.py in a python script. Ensure internet connectivity before the first run, as it downloads the MNIST dataset and saves the model as "knn.sav" or "cnn.hdf5" depending upon the modeltype variable.

  python3 Start.py