Skip to content

A four week university project where we attempted to beat Super Mario Bros. using reinforcement learning algorithms DDQN and PPO. The solution is not perfected due to time constraints, but positive results were achieved.

Notifications You must be signed in to change notification settings

sondremalerud/IDATT2502-project

Repository files navigation

Demo

Super Mario Bros. DDQN Demo

Requirements

Python

This project was developed and tested only on Python 3.11, mileage on other versions may vary based on whether dependencies such as Pytorch are available for your Python version

A C++ compiler

As the nes-py dependency bundles a NES emulator build in C++, a compiler is required (GCC recommended)

Python packages

All dependencies can be installed by running pip install -r requirements.txt from the root of the project

A copy of Super Mario Bros.

A dumped copy of Super Mario Bros. for the NES is needed for the project to be ran, and should be placed in the root of the project with file name mario.nes. The MD5 hash of our game file is 811b027eaf99c2def7b933c5208636de, yours should match to ensure compatibility.

About

A four week university project where we attempted to beat Super Mario Bros. using reinforcement learning algorithms DDQN and PPO. The solution is not perfected due to time constraints, but positive results were achieved.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •  

Languages