This is a reinforcement learning project aimed at testing TTRPG rules and monsters -- such as Open Game License (OGL) D&D5e and beasts such as Rats. The end goal is to properly create deadly monsters for dungeons and learn what behavoirs may lead to thier survival.
Examples of rules and monsters can be found D&D SRD OGL v5.1.
The approach used for this project is to create various dungeons/Gymnasium Environments and trying various methods to train agents/monsters including but not limited to Q-Learning, DQN, or other algorithms.
Adversarial ML could also become an interesting method; pitting monster against monster.
This project is implemented using Python 3.X.
All libraries used are included in the requirements.txt. You can execute the following to update your dependencies
$ pip install -r requirements.txt
Any sequential models are originally created with the Nvidia GeForce GTX 980 GPU and Intel Core i7-5960X CPU. Therefore the complexity of the models is further constrained by allowable training time and gpu memory.
Usages are still being constructed.
This project demonstrates various investigations in how one can train thier monsters.
Credit to visual assets not created by contributors of the project will go here.
giants-bane-min.jpg is by Daniel's Maps portal.gif is by phantompipa