本文整理汇总了Python中pybrain.rl.agents.LearningAgent.logging方法的典型用法代码示例。如果您正苦于以下问题：Python LearningAgent.logging方法的具体用法？Python LearningAgent.logging怎么用？Python LearningAgent.logging使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类pybrain.rl.agents.LearningAgent的用法示例。
在下文中一共展示了LearningAgent.logging方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: someEpisodes
# 需要导入模块: from pybrain.rl.agents import LearningAgent [as 别名]
# 或者: from pybrain.rl.agents.LearningAgent import logging [as 别名]
def someEpisodes(game_env, net, discountFactor=0.99, maxSteps=100, avgOver=1, returnEvents=False):
""" Return the fitness value for one episode of play, given the policy defined by a neural network. """
task = GameTask(game_env)
game_env.recordingEnabled = True
game_env.reset()
net.reset()
task.maxSteps=maxSteps
agent = LearningAgent(net)
agent.learning = False
agent.logging = False
exper = EpisodicExperiment(task, agent)
fitness = 0
for _ in range(avgOver):
rs = exper.doEpisodes(1)
# add a slight bonus for more exploration, if rewards are identical
fitness += len(set(game_env._allEvents)) * 1e-6
# the true, discounted reward
fitness += sum([sum([v*discountFactor**step for step, v in enumerate(r)]) for r in rs])
fitness /= avgOver
if returnEvents:
return fitness, game_env._allEvents
else:
return fitness