### Instructions ###

Following libraries are needed to run the project:
- Gym
- Tensorflow
- Keras

Please run the the first two cells in the  evaluation.ipynb. This will train the agents, make them play the game of connect 4 and show the game results. Deep Q-Learning with MLP will be trained on 14500 games and Deep Q-Learning with RNN will be trained on 13000 games. Then both models will play 500 games with each other and the results will be shown.