Description
Q-learning example by Valohai titled "Reinforcement Learning Tutorial Cloud Q-Learning"
Notes
Q-table is a 2 by 5 array with index 0 signifying the reward for the agent moving forward, and index 1 signifying the reward for the agent moving backwards.
[0, 0, 0, 42, 0] FORWARD
[0, 10, 0, 0, 0] BACKWARD
Execution
python3 train.py --agent=, where can be ACCOUNTANT, GAMBLER, DRUNKARD. Otherwise runs the DEEP_GAMBLER agent.
E.g.
python3 train.py --agent=DRUNKARD