pokemon-vgc-engine

To run the experiments for the SIN5021 class, go to RLStudies folder and run DQNTrainer.py and PlanningAlgorithms.py

Pokémon VGC AI Framework

(Legacy repository: https://gitlab.com/DracoStriker/simplified-pokemon-environment)

Installation

Install Python 3.6.8 or higher.
Clone this project.
Install the requirements.txt
Use you preferred Interactive Development Environment.

Alternatively you may use the Dockerfile to create a ready to run container. All dependencies are installed in the venv vgc-env and project is found in the /vgc-ai folder. User root has vgc as password. A SSH server is installed and run on the container boot.

Project Organization

In the competitor module you can store your developed agents to be used as plugins. There is a small provided example.

In the example module can be found multiple how to use the framework examples to train or test isolated agents or behaviours or run full ecosystems.

In the framework module is the core implementation of the VGC AI Framework.

In the test module is contained some unit tests from the core framework modules.

Documentation

All the documentation from Applicational Programming Interface and Framework architecture can be found in the wiki.

TODO