Tic Tac Toe solver
Trained the VGG16 model on the training images after augmenting them to increase the quantity of training images.
The input grid image is broken into 9 smaller images (of each grid) and then passed into the trained VGG16 model to classify whether X or O.
Further, the board state is passed to the minimax algorithm which predicts the probable winner from that position.
The VGG16 based model can be found at this link
Created by Team Neural Network Nerds (NNN)
Akshat Kasarwal
Joel John Mathew
Soham Mukherjee
(Wing C4)
However, due to some reason, I'm unable to run the file 'csv_file_generation.ipynb' on colab as it keeps crashing 'memory not sufficient'.