A Reinforcement Mario game using the PPO(Proximal Policy Optimization) algorithm and OpenAI gym environment.
Before running the project make sure that you have installed the following:
- Visual Studio with C++ compiler
https://visualstudio.microsoft.com/free-developer-offers/ - nes_py library
!pip install nes_py - gym library
!pip install gym - gym_super_mario_bros library
!pip install gym_super_mario_bros - stable_baselines3 library
!pip install stable_baselines3
In order to achieve a good performance you have to train the model for more than one million time steps model.learn(total_timesteps=1000000), which requires a lot of computation power and resources, so I encourage you to use cloud services to run the model, such as google Colab.