-
Notifications
You must be signed in to change notification settings - Fork 12
/
policies.py
123 lines (81 loc) · 3.72 KB
/
policies.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
##########################################################
######################## Policies ########################
##########################################################
"""
One can use two kinds of policies to do rollout of the game :
- UniformPolicy() which plays randomly from the tuple (o, r, d)
- ExpertPolicy(p_a_s) which mimics an expert that has access to the true state `s`.
One needs to provide the action distribution `p_a_s`.
In both cases, one would only need to call of the `.action()` property of the Policy class,
providing all the environment data. As a remark, it returns a one-hot encoded action,
and the gym environment requires to have an *int*-type action (just use `action.argmax()` then).
action = policy.action(o, r, done, **info)
One can collect episodes by simply calling `rollout`, given an env and a policy:
episode = rollout(env, default_policy)
"""
import torch
class Policy(torch.nn.Module):
""" Empty class, must include
1. a Reset method
3. a Action method to act in environment from 4-uplets (o, r, d, info)
"""
def reset(self):
raise NotImplemented
def action(self, o, r, d, **info):
raise NotImplemented
class UniformPolicy(Policy):
""" Uniform policy to act within the environement with random distributed actions """
def __init__(self, a_nvals):
super().__init__()
self.a_nvals = a_nvals
self.h = []
def reset(self):
self.h.clear()
def action(self, o, r, d, **info):
self.h += [o, r, d]
a = torch.distributions.one_hot_categorical.OneHotCategorical(\
probs=torch.ones(self.a_nvals)/self.a_nvals).sample(o.shape[:-1])
self.h += [a]
return a
class ExpertPolicy(Policy):
""" Expert Policy Class that chooses its actions from the hidden state s """
def __init__(self, p_a_s):
super().__init__()
self.probs_a_s = p_a_s #p(a_t | s_t, i=0)
def reset(self):
pass
def action(self, o, r, done, **info):
s_index = info["s"].argmax()
a = torch.distributions.one_hot_categorical.OneHotCategorical(probs=self.probs_a_s[s_index], ).sample()
return a
class AugmentedPolicy(Policy):
""" Augmented Policy which uses am augmented model to estimate the belived state
of the environment. It then chooses its actions according to the believed state estimate."""
def __init__(self, augmentedmodel, regime=torch.tensor(0), with_done=False):
super().__init__()
self.m = augmentedmodel
self.q_s_h = None
self.last_action = None
self.regime = regime
def update_hidden_state(self, o, r, d, with_done=False):
log_q_s_h = self.m.log_q_s_h(regime=self.regime,
loq_q_sprev_hprev=self.hidden_state,
a=self.last_action,
o=o.unsqueeze(0),
r=r.unsqueeze(0),
d=d.unsqueeze(0),
with_done=False)
self.q_s_h = torch.exp(log_q_s_h)
def reset(self):
self.q_s_h = None
self.last_action = None
def action(self, o, r, done, deterministic=False, **info):
print(self.hidden_state, self.last_action)
self.update_hidden_state(o, r, done)
new_action_p = torch.exp(self.m.log_q_a_s(a=None, s=self.q_s_h))
if deterministic :
new_action_p = new_action_p.round()
a = torch.distributions.one_hot_categorical.OneHotCategorical(
probs=new_action_p,).sample()
self.last_action = a
return a