EricKolibacz / nim

Q and DeepQ Reinforcement Learning for NIM

Q and DeepQ Reinforcement Learning for Nim

Four Nim (last player to take loses) players are available:

• Human Player (CLI inputs define the moves)
• Random Player (choosing random valid moves)
• Q Player (learning with q-table)
• DeepQ Player (learning q-table through NN)

plus some improved variants of the latter. The game can be started through the main.py-file.