def __init__(self,
policy,
*args,
observation_selector=(lambda x: x["proprio_observation"]),
reward_scale=1.0,
**kwargs):
Relabeler.__init__(self, *args, **kwargs)
self.policy = policy
self.observation_selector = observation_selector
self.reward_scale = reward_scale
def __init__(self,
*args,
achieved_goal_selector=(lambda x: x["achieved_goal"]),
goal_selector=(lambda x: x["goal"]),
goal_assigner=default_goal_assigner,
**kwargs):
Relabeler.__init__(self, *args, **kwargs)
self.achieved_goal_selector = achieved_goal_selector
self.goal_selector = goal_selector
self.goal_assigner = goal_assigner
def __init__(self,
lower_level_policy,
*args,
observation_selector=(lambda x: x["proprio_observation"]),
num_samples=8,
**kwargs):
Relabeler.__init__(self, *args, **kwargs)
self.lower_level_policy = lower_level_policy
self.observation_selector = observation_selector
self.num_samples = num_samples
def __init__(self,
*args,
observation_selector=(lambda x: x["proprio_observation"]),
order=2,
threshold=0.1,
penalty=(-1.0),
**kwargs):
Relabeler.__init__(self, *args, **kwargs)
self.observation_selector = observation_selector
self.order = order
self.threshold = threshold
self.penalty = penalty
def __init__(
self,
*args,
observation_selector=(lambda x: x["proprio_observation"]),
goal_selector=(lambda x: x["goal"]),
order=2,
reward_scale=1.0,
**kwargs
):
Relabeler.__init__(self, *args, **kwargs)
self.observation_selector = observation_selector
self.goal_selector = goal_selector
self.order = order
self.reward_scale = reward_scale
def __init__(self,
*args,
observation_selector=(lambda x: x["proprio_observation"]),
**kwargs):
Relabeler.__init__(self, *args, **kwargs)
self.observation_selector = observation_selector