-
Notifications
You must be signed in to change notification settings - Fork 5
/
generator.py
266 lines (219 loc) · 12.5 KB
/
generator.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
260
261
262
263
264
265
266
import tensorflow as tf
from tensorflow.python.ops import tensor_array_ops, control_flow_ops
class Generator(object):
def __init__(self, num_emb, batch_size, emb_dim, hidden_dim,
sequence_length, start_token,
learning_rate=0.01, reward_gamma=0.95, domain_name='gan'):
self.num_emb = num_emb
self.batch_size = batch_size
self.emb_dim = emb_dim
self.hidden_dim = hidden_dim
self.sequence_length = sequence_length
self.start_token = tf.constant([start_token] * self.batch_size, dtype=tf.int32)
self.learning_rate = tf.Variable(float(learning_rate), trainable=False)
self.reward_gamma = reward_gamma
self.g_params = []
self.d_params = []
self.temperature = 1.0
self.grad_clip = 5.0
# for tensorboard
self.g_count = 0
self.domain_name = domain_name
self.expected_reward = tf.Variable(tf.zeros([self.sequence_length]))
with tf.variable_scope(domain_name + '_generator'):
self.g_embeddings = tf.Variable(self.init_matrix([self.num_emb, self.emb_dim]), name=domain_name + '_embed')
self.g_params.append(self.g_embeddings)
self.g_recurrent_unit = self.create_recurrent_unit(self.g_params,
domain_name) # maps h_tm1 to h_t for generator
self.g_output_unit = self.create_output_unit(self.g_params,
domain_name) # maps h_t to o_t (output token logits)
# placeholder definition
self.x = tf.placeholder(tf.int32, shape=[self.batch_size,
self.sequence_length],
name=domain_name + '_gen_input_x') # sequence of tokens generated by generator
self.rewards = tf.placeholder(tf.float32, shape=[self.batch_size,
self.sequence_length],
name=domain_name + '_gen_rewards') # get from rollout policy and discriminator
# processed for batch
with tf.device("/cpu:0"):
self.processed_x = tf.transpose(tf.nn.embedding_lookup(self.g_embeddings, self.x),
perm=[1, 0, 2],
name=domain_name + '_x_batch') # seq_length x batch_size x emb_dim
# Initial states
self.h0 = tf.zeros([self.batch_size, self.hidden_dim])
self.h0 = tf.stack([self.h0, self.h0], name=domain_name + '_h0')
gen_o = tensor_array_ops.TensorArray(dtype=tf.float32, size=self.sequence_length,
dynamic_size=False, infer_shape=True, name=domain_name + '_gen_o')
gen_x = tensor_array_ops.TensorArray(dtype=tf.int32, size=self.sequence_length,
dynamic_size=False, infer_shape=True, name=domain_name + '_gen_x')
def _g_recurrence(i, x_t, h_tm1, gen_o, gen_x):
h_t = self.g_recurrent_unit(x_t, h_tm1) # hidden_memory_tuple
o_t = self.g_output_unit(h_t) # batch x vocab , logits not prob
log_prob = tf.log(tf.nn.softmax(o_t))
next_token = tf.cast(tf.reshape(tf.multinomial(log_prob, 1), [self.batch_size]), tf.int32)
x_tp1 = tf.nn.embedding_lookup(self.g_embeddings, next_token) # batch x emb_dim
gen_o = gen_o.write(i, tf.reduce_sum(tf.multiply(tf.one_hot(next_token, self.num_emb, 1.0, 0.0),
tf.nn.softmax(o_t)), 1)) # [batch_size] , prob
gen_x = gen_x.write(i, next_token) # indices, batch_size
return i + 1, x_tp1, h_t, gen_o, gen_x
_, _, _, self.gen_o, self.gen_x = control_flow_ops.while_loop(
cond=lambda i, _1, _2, _3, _4: i < self.sequence_length,
body=_g_recurrence,
loop_vars=(tf.constant(0, dtype=tf.int32),
tf.nn.embedding_lookup(self.g_embeddings, self.start_token),
self.h0,
gen_o,
gen_x),
name=domain_name + '_while_gen_o_x'
)
self.gen_x = self.gen_x.stack() # seq_length x batch_size
self.gen_x = tf.transpose(self.gen_x, perm=[1, 0],
name=domain_name + '_gen_x_stack') # batch_size x seq_length
# supervised pretraining for generator
g_predictions = tensor_array_ops.TensorArray(
dtype=tf.float32, size=self.sequence_length,
dynamic_size=False, infer_shape=True, name=domain_name + '_g_pred')
ta_emb_x = tensor_array_ops.TensorArray(
dtype=tf.float32, size=self.sequence_length, name=domain_name + '_ta_emb_x')
ta_emb_x = ta_emb_x.unstack(self.processed_x)
def _pretrain_recurrence(i, x_t, h_tm1, g_predictions):
h_t = self.g_recurrent_unit(x_t, h_tm1)
o_t = self.g_output_unit(h_t)
g_predictions = g_predictions.write(i, tf.nn.softmax(o_t)) # batch x vocab_size
x_tp1 = ta_emb_x.read(i)
return i + 1, x_tp1, h_t, g_predictions
_, _, _, self.g_predictions = control_flow_ops.while_loop(
cond=lambda i, _1, _2, _3: i < self.sequence_length,
body=_pretrain_recurrence,
loop_vars=(tf.constant(0, dtype=tf.int32),
tf.nn.embedding_lookup(self.g_embeddings, self.start_token),
self.h0, g_predictions), name=domain_name + '_while_g_pred')
self.g_predictions = tf.transpose(self.g_predictions.stack(),
perm=[1, 0, 2],
name=domain_name + '_self_g_pred') # batch_size x seq_length x vocab_size
with tf.name_scope(domain_name + '_gen_pretraining'):
# pretraining loss
self.pretrain_loss = -tf.reduce_sum(
tf.one_hot(tf.to_int32(tf.reshape(self.x, [-1])), self.num_emb, 1.0, 0.0) * tf.log(
tf.clip_by_value(tf.reshape(self.g_predictions, [-1, self.num_emb]), 1e-20, 1.0)
)
) / (self.sequence_length * self.batch_size)
self.s_pretrain_loss = tf.summary.scalar('gen_pretrain_loss', self.pretrain_loss)
# training updates
pretrain_opt = self.g_optimizer(self.learning_rate)
self.pretrain_grad, _ = tf.clip_by_global_norm(tf.gradients(self.pretrain_loss, self.g_params),
self.grad_clip)
self.pretrain_updates = pretrain_opt.apply_gradients(zip(self.pretrain_grad, self.g_params))
#######################################################################################################
# Unsupervised Training
#######################################################################################################
with tf.name_scope(domain_name + '_gen_gan_training'):
self.g_loss = -tf.reduce_sum(
tf.reduce_sum(
tf.one_hot(tf.to_int32(tf.reshape(self.x, [-1])), self.num_emb, 1.0, 0.0) *
tf.log(
tf.clip_by_value(tf.reshape(self.g_predictions, [-1, self.num_emb]), 1e-20, 1.0)
)
, 1
) * tf.reshape(self.rewards, [-1])
)
self.s_g_loss = tf.summary.scalar('gen_g_loss', self.g_loss)
g_opt = self.g_optimizer(self.learning_rate)
self.g_grad, _ = tf.clip_by_global_norm(tf.gradients(self.g_loss, self.g_params), self.grad_clip)
self.g_updates = g_opt.apply_gradients(zip(self.g_grad, self.g_params))
def generate_pretrain_summary(self, sess, x):
_summ = sess.run(
tf.summary.merge([
self.s_pretrain_loss
]),
feed_dict={
self.x: x
}
)
cur_g_count = self.g_count
#self.g_count += 1
return cur_g_count, _summ
def generate_gan_summary(self, sess, x, reward):
_summ = sess.run(
tf.summary.merge([
self.s_g_loss
]),
feed_dict={
self.x: x,
self.rewards: reward
}
)
cur_g_count = self.g_count
#self.g_count += 1
return cur_g_count, _summ
def generate(self, sess):
outputs = sess.run(self.gen_x)
return outputs
def pretrain_step(self, sess, x):
outputs = sess.run([self.pretrain_updates, self.pretrain_loss], feed_dict={self.x: x})
return outputs
def init_matrix(self, shape):
return tf.random_normal(shape, stddev=0.1)
def init_vector(self, shape):
return tf.zeros(shape)
def create_recurrent_unit(self, params, domain_name):
with tf.variable_scope(domain_name + '_rec'):
# Weights and Bias for input and hidden tensor
self.Wi = tf.Variable(self.init_matrix([self.emb_dim, self.hidden_dim]), name=domain_name + '_Wi')
self.Ui = tf.Variable(self.init_matrix([self.hidden_dim, self.hidden_dim]), name=domain_name + '_Ui')
self.bi = tf.Variable(self.init_matrix([self.hidden_dim]), name=domain_name + '_bi')
self.Wf = tf.Variable(self.init_matrix([self.emb_dim, self.hidden_dim]), name=domain_name + '_Wf')
self.Uf = tf.Variable(self.init_matrix([self.hidden_dim, self.hidden_dim]), name=domain_name + '_Uf')
self.bf = tf.Variable(self.init_matrix([self.hidden_dim]), name=domain_name + '_bf')
self.Wog = tf.Variable(self.init_matrix([self.emb_dim, self.hidden_dim]), name=domain_name + '_Wog')
self.Uog = tf.Variable(self.init_matrix([self.hidden_dim, self.hidden_dim]), name=domain_name + '_Uog')
self.bog = tf.Variable(self.init_matrix([self.hidden_dim]), name=domain_name + '_bog')
self.Wc = tf.Variable(self.init_matrix([self.emb_dim, self.hidden_dim]), name=domain_name + '_Wc')
self.Uc = tf.Variable(self.init_matrix([self.hidden_dim, self.hidden_dim]), name=domain_name + '_Uc')
self.bc = tf.Variable(self.init_matrix([self.hidden_dim]), name=domain_name + '_bc')
params.extend([
self.Wi, self.Ui, self.bi,
self.Wf, self.Uf, self.bf,
self.Wog, self.Uog, self.bog,
self.Wc, self.Uc, self.bc])
def unit(x, hidden_memory_tm1):
previous_hidden_state, c_prev = tf.unstack(hidden_memory_tm1)
# Input Gate
i = tf.sigmoid(
tf.matmul(x, self.Wi) +
tf.matmul(previous_hidden_state, self.Ui) + self.bi
)
# Forget Gate
f = tf.sigmoid(
tf.matmul(x, self.Wf) +
tf.matmul(previous_hidden_state, self.Uf) + self.bf
)
# Output Gate
o = tf.sigmoid(
tf.matmul(x, self.Wog) +
tf.matmul(previous_hidden_state, self.Uog) + self.bog
)
# New Memory Cell
c_ = tf.nn.tanh(
tf.matmul(x, self.Wc) +
tf.matmul(previous_hidden_state, self.Uc) + self.bc
)
# Final Memory cell
c = f * c_prev + i * c_
# Current Hidden state
current_hidden_state = o * tf.nn.tanh(c)
return tf.stack([current_hidden_state, c])
return unit
def create_output_unit(self, params, domain_name):
self.Wo = tf.Variable(self.init_matrix([self.hidden_dim, self.num_emb]), name=domain_name + '_Wo')
self.bo = tf.Variable(self.init_matrix([self.num_emb]), name=domain_name + '_bo')
params.extend([self.Wo, self.bo])
def unit(hidden_memory_tuple):
hidden_state, c_prev = tf.unstack(hidden_memory_tuple)
# hidden_state : batch x hidden_dim
logits = tf.matmul(hidden_state, self.Wo) + self.bo
# output = tf.nn.softmax(logits)
return logits
return unit
def g_optimizer(self, *args, **kwargs):
return tf.train.AdamOptimizer(*args, **kwargs)