-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathhashed_mem_nw.py
192 lines (163 loc) · 9.32 KB
/
hashed_mem_nw.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
"""Description:- Hashed Memory Networks with BoW and GRU reader."""
from __future__ import absolute_import
from __future__ import division
import tensorflow as tf
from six.moves import range
import numpy as np
# Implementing position encoding.
def position_encoding(sentence_size, embedding_size):
encoding = np.ones((embedding_size, sentence_size), dtype=np.float32)
ls = sentence_size+1
le = embedding_size+1
for i in range(1, le):
for j in range(1, ls):
encoding[i-1, j-1] = (i - (le-1)/2) * (j - (ls-1)/2)
encoding = 1 + 4 * encoding / embedding_size / sentence_size
return np.transpose(encoding)
""" Adding gradient noise with the input Tensor `t` as a gradient and the output tensor will be `t` + gaussian noise."""
def add_gradient_noise(t, stddev=1e-3, name=None):
with tf.op_scope([t, stddev], name, "add_gradient_noise") as name:
t = tf.convert_to_tensor(t, name="t")
gn = tf.random_normal(tf.shape(t), stddev=stddev)
return tf.add(t, gn, name=name)
"""Overwriting the nil_slot (first row) of the input Tensor with zeros. As it is not trained."""
def zero_nil_slot(t, name=None):
with tf.op_scope([t], name, "zero_nil_slot") as name:
t = tf.convert_to_tensor(t, name="t")
s = tf.shape(t)[1]
z = tf.zeros(tf.pack([1, s]))
return tf.concat(0,[z, tf.slice(t, [1, 0], [-1, -1])], name=name)
"""Key Value Memory Network."""
class Hashed_Mem_Nw(object):
def __init__(self, vocab_size,query_size, story_size, memory_key_size,
memory_value_size, embedding_size, reader='bow',
l2_lambda=0.2):
#Initializing the current object with parameters.
feature_size = 40 #defalt value
self._batch_size = 32
self._hops = 3
self._name = 'KeyValueMemN2N'
self._story_size = story_size
self._vocab_size = vocab_size
self._query_size = query_size
self._memory_key_size = memory_key_size
self._embedding_size = embedding_size
self._memory_value_size = memory_value_size
self._encoding = tf.constant(position_encoding(self._story_size, self._embedding_size), name="encoding")
self._reader = reader
self._build_inputs()
d = feature_size
self._feature_size = feature_size
self._n_hidden = feature_size
self.reader_feature_size = 0
# trainable variables
if reader == 'bow':
self.reader_feature_size = self._embedding_size
elif reader == 'simple_gru':
self.reader_feature_size = self._n_hidden
self.A = tf.get_variable('A', shape=[self._feature_size, self.reader_feature_size],
initializer=tf.contrib.layers.xavier_initializer())
self.TK = tf.get_variable('TK', shape=[self._memory_value_size, self.reader_feature_size],
initializer=tf.contrib.layers.xavier_initializer())
self.TV = tf.get_variable('TV', shape=[self._memory_value_size, self.reader_feature_size],
initializer=tf.contrib.layers.xavier_initializer())
# Embedding layers and placeholders for tensorflow.
with tf.device('/cpu:0'), tf.name_scope("embedding"):
nil_word_slot = tf.zeros([1, embedding_size])
self.W = tf.concat(0, [nil_word_slot, tf.get_variable('W', shape=[vocab_size-1, embedding_size],
initializer=tf.contrib.layers.xavier_initializer())])
self.W_memory = tf.concat(0,[nil_word_slot, tf.get_variable('W_memory', shape=[vocab_size-1, embedding_size],
initializer=tf.contrib.layers.xavier_initializer())])
self._nil_vars = set([self.W.name, self.W_memory.name])
self.embedded_chars = tf.nn.embedding_lookup(self.W, self._query)
self.mkeys_embedded_chars = tf.nn.embedding_lookup(self.W_memory, self._memory_key)
self.mvalues_embedded_chars = tf.nn.embedding_lookup(self.W_memory, self._memory_value)
if reader == 'bow':
q_r = tf.reduce_sum(self.embedded_chars*self._encoding, 1)
doc_r = tf.reduce_sum(self.mkeys_embedded_chars*self._encoding, 2)
value_r = tf.reduce_sum(self.mvalues_embedded_chars*self._encoding, 2)
elif reader == 'simple_gru':
x_tmp = tf.reshape(self.mkeys_embedded_chars, [-1, self._story_size, self._embedding_size])
x = tf.transpose(x_tmp, [1, 0, 2])
# Reshape to (n_steps*batch_size, n_input)
x = tf.reshape(x, [-1, self._embedding_size])
# Split to get a list of 'n_steps'
x = tf.split(0, self._story_size, x)
# splitting the questions
q = tf.transpose(self.embedded_chars, [1, 0, 2])
q = tf.reshape(q, [-1, self._embedding_size])
q = tf.split(0, self._query_size, q)
k_rnn = tf.nn.rnn_cell.GRUCell(self._n_hidden)
q_rnn = tf.nn.rnn_cell.GRUCell(self._n_hidden)
with tf.variable_scope('story_gru'):
doc_output, _ = tf.nn.rnn(k_rnn, x, dtype=tf.float32)
with tf.variable_scope('question_gru'):
q_output, _ = tf.nn.rnn(q_rnn, q, dtype=tf.float32)
doc_r = tf.nn.dropout(tf.reshape(doc_output[-1], [-1, self._memory_key_size, self._n_hidden]), self.keep_prob)
value_r = doc_r
q_r = tf.nn.dropout(q_output[-1], self.keep_prob)
r_list = []
for _ in range(self._hops):
# define R for variables
R = tf.get_variable('R{}'.format(_), shape=[self._feature_size, self._feature_size],
initializer=tf.contrib.layers.xavier_initializer())
r_list.append(R)
o = self._key_addressing(doc_r, value_r, q_r, r_list)
o = tf.transpose(o)
if reader == 'bow':
self.B = self.A
elif reader == 'simple_gru':
self.B = tf.get_variable('B', shape=[self._feature_size, self._embedding_size],
initializer=tf.contrib.layers.xavier_initializer())
y_tmp = tf.matmul(self.B, self.W_memory, transpose_b=True)
with tf.name_scope("prediction"):
logits = tf.matmul(o, y_tmp)
probs = tf.nn.softmax(tf.cast(logits, tf.float32))
# Calculating the cross entropy.
cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits, tf.cast(self._labels, tf.float32), name='cross_entropy')
cross_entropy_sum = tf.reduce_sum(cross_entropy, name="cross_entropy_sum")
# loss calculation.
vars = tf.trainable_variables()
lossL2 = tf.add_n([tf.nn.l2_loss(v) for v in vars])
loss_op = cross_entropy_sum + l2_lambda*lossL2
# predicting output
predict_op = tf.argmax(probs, 1, name="predict_op")
# assign outputs
self.loss_op = loss_op
self.predict_op = predict_op
self.probs = probs
#Below method declares the place holders.
def _build_inputs(self):
with tf.name_scope("input"):
self._memory_key = tf.placeholder(tf.int32, [None, self._memory_value_size, self._story_size], name='memory_key')
self._query = tf.placeholder(tf.int32, [None, self._query_size], name='question')
self._memory_value = tf.placeholder(tf.int32, [None, self._memory_value_size, self._story_size], name='memory_value')
self._labels = tf.placeholder(tf.float32, [None, self._vocab_size], name='answer')
self.keep_prob = tf.placeholder(tf.float32, name='keep_prob')
'''Below method defines the vector representation for keys in memory where each memory key is of dimension [1, embedding_size]
and the vector representation for values in memory where each values has a dimension of [1, embedding_size] and the vector representation
for the question with dimension as [1, embedding_size]'''
def _key_addressing(self, mkeys, mvalues, questions, r_list):
with tf.variable_scope(self._name):
u = tf.matmul(self.A, questions, transpose_b=True)
u = [u]
for _ in range(self._hops):
R = r_list[_]
u_temp = u[-1]
mk_temp = mkeys + self.TK
k_temp = tf.reshape(tf.transpose(mk_temp, [2, 0, 1]), [self.reader_feature_size, -1])
a_k_temp = tf.matmul(self.A, k_temp)
a_k = tf.reshape(tf.transpose(a_k_temp), [-1, self._memory_key_size, self._feature_size])
u_expanded = tf.expand_dims(tf.transpose(u_temp), [1])
dotted = tf.reduce_sum(a_k*u_expanded, 2)
probs = tf.nn.softmax(dotted)
probs_expand = tf.expand_dims(probs, -1)
mv_temp = mvalues + self.TV
v_temp = tf.reshape(tf.transpose(mv_temp, [2, 0, 1]), [self.reader_feature_size, -1])
a_v_temp = tf.matmul(self.A, v_temp)
a_v = tf.reshape(tf.transpose(a_v_temp), [-1, self._memory_key_size, self._feature_size])
o_k = tf.reduce_sum(probs_expand*a_v, 1)
o_k = tf.transpose(o_k)
u_k = tf.matmul(R, u[-1]+o_k)
u.append(u_k)
return u[-1]