https://github.com/dvorjackz/MarioRL

Deep double Q-learning to help Mario beat Super Mario Bros on SNES
https://github.com/dvorjackz/MarioRL

double-q- python reinforcement-learning tensorflow

Last synced: 3 months ago
JSON representation

Deep double Q-learning to help Mario beat Super Mario Bros on SNES

• Host: GitHub
• URL: https://github.com/dvorjackz/MarioRL
• Owner: dvorjackz
• Created: 2020-05-25T01:19:23.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2023-03-24T23:46:48.000Z (over 1 year ago)
• Last Synced: 2024-04-27T08:36:21.958Z (7 months ago)
• Topics: double-q-, python, reinforcement-learning, tensorflow
• Language: Python
• Homepage:
• Size: 18.5 MB
• Stars: 5
• Watchers: 4
• Forks: 0
• Open Issues: 6
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# MarioRL

Using [Gym Super Mario Bros](https://pypi.org/project/gym-super-mario-bros/) as the environment and [Stable Baselines](https://github.com/hill-a/stable-baselines), a fork of OpenAI's popular [Baselines](https://github.com/openai/baselines) reinforcement learning library, we apply concepts highlighted in recent influential papers in the RL space to traing an agent to beat Super Mario Bros for NES as quickly as possible. 

The final report on our findings is included in the repo as paper.pdf. 

## Setup

*Important*: must use Python version < 3.8, preferrably Python-3.7.6

```

python -m venv .env

source .env/bin/activate

pip install -r requirements.txt

```

## Potential PMIX Error Fix

Solve by using the following line:

```

export PMIX_MCA_gds=hash

```