Ten-Thousand / Zehntausend

First try at reinforcement leaning with Q-learning for the dice game Ten-Thousand (Zehntausend).

Examples

Playing the game interactively

python ./tenk/game.py

Train/Watch single model AI

Train and watch the ai play the game for an AI that processes all states and actions in a single dictionary.

python ./tenk/ai/single.py

Train/Watch split model AI

Train and watch the ai play the game for an AI that processes all states and actions separately for choosing dices and choosing to re-roll.

python ./tenk/ai/split.py