All-In is a poker AI developed for a bachelor thesis by Jens Kaminsky and Anya Liebendörfer.
The accompanying paper can be found here.
This code was written specifically for that paper and is neither particularly maintained
nor documented. The proper way to launch a simulation is
to execute runner.py after tuning it to desire. All parameters and data about agents
are stored in savefiles/ except for agent weights which are stored in models/. Plots and strategy
vectors can be computed by executing plot_existing_strategies.py and agent_aggregator.py
after modification and after agents have been trained.
This repository was seeded off of PyPokerEngine Competition Starter Kit and heavily uses code from SpinningUp. We are grateful for both, and all credit for their respective code goes to them.