コード例 #1
0
def use_model():
    cart_pole_ctrl = Controller(None,
                                'CartPole-v1',
                                StateBuilderCartPole(),
                                communicate=False)
    # cart_pole_ctrl = Controller(None, 'Taxi-v2', None)
    # cart_pole_ctrl = Controller(None, 'LunarLander-v2', state_builder=StateBuilderLunarLander(), communicate=False)

    learner = QLearner(cart_pole_ctrl.get_action_space(),
                       epsilon=0.0,
                       init_alpha=.5,
                       gamma=.9)

    cart_pole_ctrl.set_learner(learner)
    cart_pole_ctrl.load("models/CartPole-v1-7.model")

    count = 0
    while True:
        cart_pole_ctrl.run(render=True)
        count += 1
        print("Epoch {}".format(count))