Skip to content

Latest commit

 

History

History

nim-a2c

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
agent.py
agent.py
 
 
illustration.svg
illustration.svg
 
 
network.py
network.py
 
 
nim.py
nim.py
 
 
train.ipynb
train.ipynb
 
 

README.md

Multi-agent A2C for Nim-21

Q-values heatmap for (S, A)-combinations.

Two agents that learn to play Nim-21 using PyTorch and A2C. Both agents learn the game-theory optimal strategy. The critic values show which states are winning and losing.