This repository has been archived by the owner on Oct 31, 2023. It is now read-only.
/
act_group.py
259 lines (236 loc) · 8.3 KB
/
act_group.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
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
# Copyright (c) Facebook, Inc. and its affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
#
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
#
import set_path
set_path.append_sys_path()
import rela
import hanalearn
assert rela.__file__.endswith(".so")
assert hanalearn.__file__.endswith(".so")
class ActGroup:
def __init__(
self,
devices,
agent,
seed,
num_thread,
num_game_per_thread,
num_player,
explore_eps,
boltzmann_t,
method,
sad,
shuffle_color,
hide_action,
trinary,
replay_buffer,
multi_step,
max_len,
gamma,
off_belief,
belief_model,
):
self.devices = devices.split(",")
self.model_runners = []
methods = {"act": 5000, "compute_priority": 100}
if off_belief:
methods["compute_target"] = 5000
make_model_runners(agent, self.devices, self.model_runners, methods)
self.num_runners = len(self.model_runners)
self.off_belief = off_belief
self.belief_model = belief_model
self.belief_runner = None
if belief_model is not None:
self.belief_runner = []
for bm in belief_model:
print("add belief model to: ", bm.device)
self.belief_runner.append(
rela.BatchRunner(bm, bm.device, 5000, ["sample"])
)
self.actors = []
if method == "vdn":
for i in range(num_thread):
thread_actors = []
for j in range(num_game_per_thread):
actor = hanalearn.R2D2Actor(
self.model_runners[i % self.num_runners],
seed,
num_player,
0,
explore_eps,
boltzmann_t,
True,
sad,
shuffle_color,
hide_action,
trinary,
replay_buffer,
multi_step,
max_len,
gamma,
)
seed += 1
thread_actors.append([actor])
self.actors.append(thread_actors)
elif method == "iql":
for i in range(num_thread):
thread_actors = []
for j in range(num_game_per_thread):
game_actors = []
for k in range(num_player):
actor = hanalearn.R2D2Actor(
self.model_runners[i % self.num_runners],
seed,
num_player,
k,
explore_eps,
boltzmann_t,
False,
sad,
shuffle_color,
hide_action,
trinary,
replay_buffer,
multi_step,
max_len,
gamma,
)
if self.off_belief:
if self.belief_runner is None:
actor.set_belief_runner(None)
else:
actor.set_belief_runner(
self.belief_runner[i % len(self.belief_runner)]
)
seed += 1
game_actors.append(actor)
for k in range(num_player):
partners = game_actors[:]
partners[k] = None
game_actors[k].set_partners(partners)
thread_actors.append(game_actors)
self.actors.append(thread_actors)
print("ActGroup created")
def start(self):
for runner in self.model_runners:
runner.start()
if self.belief_runner is not None:
for runner in self.belief_runner:
runner.start()
def update_model(self, agent):
for runner in self.model_runners:
runner.update_model(agent)
class BRActGroup:
"""
Best response act group
"""
def __init__(
self,
devices,
agent,
coop_agents,
seed,
num_thread,
num_game_per_thread,
num_player,
explore_eps,
boltzmann_t,
method,
sad,
shuffle_color,
hide_action,
trinary,
replay_buffer,
multi_step,
max_len,
gamma,
):
assert method in ["iql"]
self.devices = devices.split(",")
self.model_runners = []
methods = {"act": 5000, "compute_priority": 100}
make_model_runners(agent, self.devices, self.model_runners, methods)
self.num_runners = len(self.model_runners)
self.coop_model_runners = []
if coop_agents is not None:
make_model_runners(
coop_agents, self.devices, self.coop_model_runners, methods
)
self.num_coop_runners = len(self.coop_model_runners)
print(
f"Making a best response with: {self.num_coop_runners} independent cooperative agents"
)
self.actors = []
for i in range(num_thread):
thread_actors = []
for j in range(num_game_per_thread):
game_actors = []
for k in range(num_player):
cur_explore_eps = explore_eps
if k == 0:
cur_model = self.model_runners[i % self.num_runners]
player_buffer = replay_buffer
else:
if self.coop_model_runners:
cur_model = self.coop_model_runners[
j % self.num_coop_runners
]
cur_explore_eps = [0.0]
else:
cur_model = self.model_runners[i % self.num_runners]
cur_explore_eps = [1.0]
player_buffer = None
actor = hanalearn.R2D2Actor(
cur_model,
seed,
num_player,
k,
cur_explore_eps,
boltzmann_t,
False,
sad,
shuffle_color,
hide_action,
trinary,
player_buffer,
multi_step,
max_len,
gamma,
)
seed += 1
game_actors.append(actor)
for k in range(num_player):
partners = game_actors[:]
partners[k] = None
game_actors[k].set_partners(partners)
thread_actors.append(game_actors)
self.actors.append(thread_actors)
print("ActGroup created")
def start(self):
for runner in self.model_runners:
runner.start()
for runner in self.coop_model_runners:
runner.start()
def update_model(self, agent):
for runner in self.model_runners:
runner.update_model(agent)
def update_coop_models(self, coop_agents):
for i, runner in enumerate(self.coop_model_runners):
runner.update_model(coop_agents[i % self.num_coop_runners])
def make_model_runners(agents, devices, runners, methods):
if not isinstance(agents, list):
agents = [agents]
for dev in devices:
for agent in agents:
runner = rela.BatchRunner(agent.clone(dev), dev)
for method, sample_limit in methods.items():
runner.add_method(method, sample_limit)
runners.append(runner)