hungnm2008 / double-dqn-super-mario-bros

Training a Double Deep Q Network (Double DQN) agent to play Super Mario Bros

The purpose of this work is to implement a Double DQN agent and evaluate its performance in a famous game: Super Mario Bros. We will be using the gym-super-mario-bros environment, which is built on top of the OpenAI gym library.

Note: Run the code with GPU