decentgradient / 3dtictactoe Goto Github PK

Directory List
    backupscripts - experiments done during development imports may not resolve unless moved back to parent
    chk - checkpoints - not used
    common -
        1 benchmark.py - benchmarking objects
        2 dnn.py - deep version of ddpg network and training objects
        3 game.py - the game environment
        4 nn.py - shallow version of dnn.py originally used
        5 q.py - implementation of DQN
        6 replay_buffer.py - memory buffer for dnn
    data - pickled results of experiments
    plots - images generated by pyplot of pickeld data
    results - NA

Scripts in parent directory
    pgqvrand.py - ddpg player 1 vs random player 2
    randvpgq.py   random player 1 vs ddpg player 2
    pgvpgrand.py - ddpg player 1 vs ddpg player 2 with 50% random moves
    pgrandvpg.py - ddpg player with 50% random moves  1 vs ddpg player 2