DeepQLearning

Implementation of Deep Mind's Paper titled: "Human Level Control through Deep Reinforcement Learning". The policy gradient algorithm was also implemented and tested on Cart Pole just for experiments.

Requirements

• Python 3.6.1
• Tensorflow
• Gym
• numpy
• sklearn
• matplotlib

Training the model

• For DQN python main.py When prompted to select train, play or visualize type 'train'
• For Policy Gradients python policy_gradients.py When prompted to select train or play type 'train'

Testing the model

• For DQN python main.py When prompted to select train, play or visualize type 'play'
• For Policy Gradients python policy_gradients.py When prompted to select train or play type 'play'

Results

Game	Highest Score
Breakout	45
Cart Pole	200

Breakout

Cartpole

Experiments

Breakout

Cartpole