-
Notifications
You must be signed in to change notification settings - Fork 10
/
Copy pathnew_pretrained_actor_critic_network.py
executable file
·163 lines (130 loc) · 7.73 KB
/
new_pretrained_actor_critic_network.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
import tensorflow as tf
import numpy as np
import math
# Hyper Parameters
LAYER1_SIZE = 128
LAYER2_SIZE = 64
LEARNING_RATE = 1e-4
TAU = 0.001
L2 = 0.01
BATCH_SIZE = 64
class ActorCriticNetwork:
"""docstring for ActorNetwork"""
def __init__(self,sess,state_dim,action_dim):
self.time_step = 0
self.sess = sess
self.state_dim = state_dim
self.action_dim = action_dim
# create eval network
self.state_input = tf.placeholder(tf.float32, shape=[None, state_dim]) # None * state_space
self.action, self.actor_net \
= self.create_eval_actor_network(self.state_input, state_dim, action_dim)
self.q_value, self.critic_net \
= self.create_eval_critic_network(self.state_input, self.action, state_dim, action_dim)
# create target network
self.ema = tf.train.ExponentialMovingAverage(decay=1-TAU)
self.target_state_input = tf.placeholder(tf.float32, shape=[None, state_dim]) # None * state_space
self.target_action, self.target_actor_net, self.target_actor_update,\
= self.create_target_actor_network(self.ema, self.target_state_input, self.actor_net)
self.target_q_value,self.target_critic_net,self.target_critic_update \
= self.create_target_critic_network(self.ema, self.target_state_input, self.target_action, self.critic_net)
# define training rules
self.create_training_actor_method(self.actor_net)
self.create_training_critic_method(self.critic_net)
self.sess.run(tf.initialize_all_variables())
self.update_target()
def create_training_critic_method(self, net):
self.y_input = tf.placeholder("float",[None,1])
weight_decay = tf.add_n([L2 * tf.nn.l2_loss(var) for var in net])
self.cost = tf.reduce_mean(tf.square(self.y_input - self.q_value)) + weight_decay
self.critic_optimizer = tf.train.AdamOptimizer(LEARNING_RATE).\
minimize(self.cost, var_list=self.critic_net)
def create_training_actor_method(self, net):
self.actor_loss = -tf.reduce_mean(self.q_value)
self.actor_optimizer = tf.train.AdamOptimizer(LEARNING_RATE).\
minimize(self.actor_loss, var_list=self.actor_net)
def create_eval_actor_network(self, state_embed_input, state_dim, action_dim):
layer1_size = LAYER1_SIZE
layer2_size = LAYER2_SIZE
W1 = self.variable([state_dim,layer1_size],state_dim)
b1 = self.variable([layer1_size],state_dim)
W2 = self.variable([layer1_size,layer2_size],layer1_size)
b2 = self.variable([layer2_size],layer1_size)
W3 = tf.Variable(tf.random_uniform([layer2_size,action_dim],-3e-3,3e-3))
b3 = tf.Variable(tf.random_uniform([action_dim],-3e-3,3e-3))
layer1 = tf.nn.relu(tf.matmul(state_embed_input,W1) + b1)
layer2 = tf.nn.relu(tf.matmul(layer1,W2) + b2)
action_output = tf.tanh(tf.matmul(layer2,W3) + b3)
return action_output, [W1,b1,W2,b2,W3,b3]
def create_eval_critic_network(self,state_embed_input, action_input, state_dim, action_dim):
layer1_size = LAYER1_SIZE
layer2_size = LAYER2_SIZE
W1 = self.variable([state_dim,layer1_size],state_dim)
b1 = self.variable([layer1_size],state_dim)
W2 = self.variable([layer1_size,layer2_size],layer1_size+action_dim)
W2_action = self.variable([action_dim,layer2_size],layer1_size+action_dim)
b2 = self.variable([layer2_size],layer1_size+action_dim)
W3 = tf.Variable(tf.random_uniform([layer2_size,1],-3e-3,3e-3))
b3 = tf.Variable(tf.random_uniform([1],-3e-3,3e-3))
layer1 = tf.nn.relu(tf.matmul(state_embed_input,W1) + b1)
layer2 = tf.nn.relu(tf.matmul(layer1,W2) + tf.matmul(action_input,W2_action) + b2)
q_value_output = tf.identity(tf.matmul(layer2,W3) + b3)
return q_value_output,[W1,b1,W2,W2_action,b2,W3,b3]
def create_target_actor_network(self, ema, state_embed_input, actor_net):
target_update = ema.apply(actor_net)
target_net = [ema.average(x) for x in actor_net]
layer1 = tf.nn.relu(tf.matmul(state_embed_input,target_net[0]) + target_net[1])
layer2 = tf.nn.relu(tf.matmul(layer1,target_net[2]) + target_net[3])
action_output = tf.tanh(tf.matmul(layer2,target_net[4]) + target_net[5])
return action_output, target_net, target_update
def create_target_critic_network(self, ema, state_embed_input, action_input, critic_net):
target_update = ema.apply(critic_net)
target_net = [ema.average(x) for x in critic_net]
layer1 = tf.nn.relu(tf.matmul(state_embed_input,target_net[0]) + target_net[1])
layer2 = tf.nn.relu(tf.matmul(layer1,target_net[2]) + tf.matmul(action_input,target_net[3]) + target_net[4])
q_value_output = tf.identity(tf.matmul(layer2,target_net[5]) + target_net[6])
return q_value_output, target_net, target_update
def update_target(self):
self.sess.run([self.target_actor_update, self.target_critic_update])
def train_critic(self,y_batch,state_batch,action_batch):
self.time_step += 1
cost,_ = self.sess.run([self.cost,self.critic_optimizer],feed_dict={
self.y_input:y_batch,
self.state_input:state_batch,
self.action:action_batch
})
return cost
def train_actor(self,state_batch):
self.sess.run(self.actor_optimizer,feed_dict={
self.state_input:state_batch
})
''' critic net '''
def target_q(self,state_batch):
return self.sess.run(self.target_q_value,feed_dict={
self.target_state_input:state_batch,
})
def q_value(self,state_batch,action_batch):
return self.sess.run(self.q_value,feed_dict={
self.state_input:state_batch,
self.action:action_batch})
''' actor net '''
def actions(self,state_batch):
return self.sess.run(self.action,feed_dict={
self.state_input:state_batch
})
# f fan-in size
def variable(self,shape,f):
return tf.Variable(tf.random_uniform(shape,-1/math.sqrt(f),1/math.sqrt(f)))
'''
def load_network(self):
self.saver = tf.train.Saver()
checkpoint = tf.train.get_checkpoint_state("saved_actor_networks")
if checkpoint and checkpoint.model_checkpoint_path:
self.saver.restore(self.sess, checkpoint.model_checkpoint_path)
print "Successfully loaded:", checkpoint.model_checkpoint_path
else:
print "Could not find old network weights"
def save_network(self,time_step):
print 'save actor-network...',time_step
self.saver.save(self.sess, 'saved_actor_networks/' + 'actor-network', global_step = time_step)
'''