tabular_q_agent.py 文件源码

python
阅读 21 收藏 0 点赞 0 评论 0

项目:deep-rl 作者: xinghai-sun 项目源码 文件源码
def __init__(self,
                 action_space,
                 observation_space,
                 q_init=0.0,
                 learning_rate=0.1,
                 discount=1.0,
                 epsilon=0.05):
        if not isinstance(action_space, spaces.Discrete):
            raise TypeError("Action space type should be Discrete.")
        if not isinstance(observation_space, spaces.Discrete):
            raise TypeError("Observation space type should be Discrete.")
        self._action_space = action_space
        self._learning_rate = learning_rate
        self._discount = discount
        self._epsilon = epsilon
        self._q = defaultdict(lambda: q_init * np.ones(action_space.n))
评论列表
文章目录


问题


面经


文章

微信
公众号

扫码关注公众号