本文整理汇总了Python中rllab.misc.special.weighted_sample方法的典型用法代码示例。如果您正苦于以下问题：Python special.weighted_sample方法的具体用法？Python special.weighted_sample怎么用？Python special.weighted_sample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rllab.misc.special的用法示例。
在下文中一共展示了special.weighted_sample方法的3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: get_action
# 需要导入模块: from rllab.misc import special [as 别名]
# 或者: from rllab.misc.special import weighted_sample [as 别名]
def get_action(self, observation):
if self.state_include_action:
if self.prev_action is None:
prev_action = np.zeros((self.action_space.flat_dim,))
else:
prev_action = self.action_space.flatten(self.prev_action)
all_input = np.concatenate([
self.observation_space.flatten(observation),
prev_action
])
else:
all_input = self.observation_space.flatten(observation)
# should not be used
prev_action = np.nan
probs, hidden_vec = [x[0] for x in self.f_step_prob([all_input], [self.prev_hidden])]
action = special.weighted_sample(probs, range(self.action_space.n))
self.prev_action = action
self.prev_hidden = hidden_vec
agent_info = dict(prob=probs)
if self.state_include_action:
agent_info["prev_action"] = prev_action
return action, agent_info 示例2: get_action
# 需要导入模块: from rllab.misc import special [as 别名]
# 或者: from rllab.misc.special import weighted_sample [as 别名]
def get_action(self, observation):
if self.state_include_action:
if self.prev_action is None:
prev_action = np.zeros((self.action_space.flat_dim,))
else:
prev_action = self.action_space.flatten(self.prev_action)
all_input = np.concatenate([
self.observation_space.flatten(observation),
prev_action
])
else:
all_input = self.observation_space.flatten(observation)
# should not be used
prev_action = np.nan
probs, hidden_vec = [x[0] for x in self.f_step_prob(
[all_input], [self.prev_hidden])]
action = special.weighted_sample(probs, range(self.action_space.n))
self.prev_action = action
self.prev_hidden = hidden_vec
agent_info = dict(prob=probs)
if self.state_include_action:
agent_info["prev_action"] = prev_action
return action, agent_info 示例3: weighted_sample
# 需要导入模块: from rllab.misc import special [as 别名]
# 或者: from rllab.misc.special import weighted_sample [as 别名]
def weighted_sample(self, weights):
return special.weighted_sample(weights, range(self.n))