-
Notifications
You must be signed in to change notification settings - Fork 21
/
transformer_keras_tensorflow.py
200 lines (162 loc) · 8.25 KB
/
transformer_keras_tensorflow.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
import keras
import tensorflow as tf
import numpy as np
def get_angles(pos, i, d_model):
angle_rates = 1 / np.power(10000, (2 * (i//2)) / np.float32(d_model))
return pos * angle_rates
def positional_encoding(position, d_model):
angle_rads = get_angles(np.arange(position)[:, np.newaxis],
np.arange(d_model)[np.newaxis, :],
d_model)
sines = np.sin(angle_rads[:, 0::2])
cosines = np.cos(angle_rads[:, 1::2])
pos_encoding = np.concatenate([sines, cosines], axis=-1)
pos_encoding = pos_encoding[np.newaxis, ...]
return tf.cast(pos_encoding, dtype=tf.float32)
def create_padding_mask(seq,seq_length):
return tf.cast(tf.sequence_mask(seq.shape[0], maxlen=seq_length), dtype = tf.int32 )
def create_look_ahead_mask(size):
mask = tf.linalg.band_part(tf.ones((size, size)), -1, 0)
return mask # (seq_len, seq_len)
def scaled_dot_product_attention(q, k, v, mask=None):
matmul_qk = tf.matmul(q, k, transpose_b=True) # (..., seq_len_q, seq_len_k)
dk = tf.cast(tf.shape(k)[-1], tf.float32)
scaled_attention_logits = matmul_qk / tf.sqrt(dk)
if mask is not None:
#mask is the relevant values from k to keep , [q.size , k .size]
scaled_attention_logits += tf.log(tf.to_float(mask))
attention_weights = tf.nn.softmax(scaled_attention_logits, axis=-1) # (..., seq_len_q, seq_len_k)
output = tf.matmul(attention_weights, v) # (..., seq_len_q, depth_v)
return output, attention_weights
class MultiHeadAttention(tf.keras.layers.Layer):
def __init__(self, d_model, num_heads):
super(MultiHeadAttention, self).__init__()
self.num_heads = num_heads
self.d_model = d_model
assert d_model % self.num_heads == 0
self.depth = d_model // self.num_heads
self.wq = tf.keras.layers.Dense(d_model)
self.wk = tf.keras.layers.Dense(d_model)
self.wv = tf.keras.layers.Dense(d_model)
self.dense = tf.keras.layers.Dense(d_model)
def split_heads(self, x):
x = tf.reshape(x, (-1, self.num_heads, self.depth))
return tf.transpose(x, perm=[1, 0, 2])
def call(self, v, k, q, mask):
q = self.wq(q) # (seq_len, d_model)
k = self.wk(k) # (seq_len, d_model)
v = self.wv(v) # (seq_len, d_model)
q = self.split_heads(q) # (num_heads, seq_len_q, depth)
k = self.split_heads(k) # (num_heads, seq_len_k, depth)
v = self.split_heads(v) # (num_heads, seq_len_v, depth)
# scaled_attention.shape == ( num_heads, seq_len_v, depth)
# attention_weights.shape == ( num_heads, seq_len_q, seq_len_k)
scaled_attention, attention_weights = scaled_dot_product_attention(q, k, v, mask)
scaled_attention = tf.transpose(scaled_attention, perm=[1, 0, 2]) # (seq_len_q, num_heads, depth)
concat_attention = tf.reshape(scaled_attention, (-1, self.d_model)) # (seq_len_q, d_model)
output = self.dense(concat_attention) # (seq_len_q, d_model)
return output, attention_weights
def point_wise_feed_forward_network(d_model, dff):
return tf.keras.Sequential([
tf.keras.layers.Dense(dff, activation='relu'), # (seq_len, dff)
tf.keras.layers.Dense(d_model) # ( seq_len, d_model)
])
class EncoderLayer(tf.keras.layers.Layer):
def __init__(self, d_model, num_heads, dff, rate=0.1):
super(EncoderLayer, self).__init__()
self.mha = MultiHeadAttention(d_model, num_heads)
self.ffn = point_wise_feed_forward_network(d_model, dff)
self.dropout1 = tf.nn.dropout(input_ , 1-rate)
self.dropout2 = tf.nn.dropout(input_ , 1-rate)
def call(self, x, training, mask):
attn_output, _ = self.mha(x, x, x, mask) # ( input_seq_len, d_model)
attn_output = self.dropout1(attn_output, training=training)
out1= tf.contrib.layers.layer_norm(x + attn_output)
ffn_output = self.ffn(out1) # ( input_seq_len, d_model)
ffn_output = self.dropout2(ffn_output, training=training)
out2 = tf.contrib.layers.layer_norm(out1 + ffn_output)
return out2
class DecoderLayer(tf.keras.layers.Layer):
def __init__(self, d_model, num_heads, dff, rate=0.1):
super(DecoderLayer, self).__init__()
self.mha1 = MultiHeadAttention(d_model, num_heads)
self.mha2 = MultiHeadAttention(d_model, num_heads)
self.ffn = point_wise_feed_forward_network(d_model, dff)
self.dropout1 = tf.nn.dropout(input_ , 1-rate)
self.dropout2 = tf.nn.dropout(input_ , 1-rate)
self.dropout3 = tf.nn.dropout(input_ , 1-rate)
def call(self, x, enc_output, training,
look_ahead_mask, padding_mask):
# enc_output.shape == (batch_size, input_seq_len, d_model)
attn1, attn_weights_block1 = self.mha1(x, x, x, look_ahead_mask) # ( target_seq_len, d_model)
attn1 = self.dropout1(attn1, training=training)
out1 = tf.contrib.layers.layer_norm(x + attn1)
attn2, attn_weights_block2 = self.mha2(
enc_output, enc_output, out1, padding_mask) # ( target_seq_len, d_model)
attn2 = self.dropout2(attn2, training=training)
out2 = tf.contrib.layers.layer_norm(attn2 + out1)
ffn_output = self.ffn(out2) # (batch_size, target_seq_len, d_model)
ffn_output = self.dropout3(ffn_output, training=training)
out3 = tf.contrib.layers.layer_norm(ffn_output + out2)
return out3, attn_weights_block1, attn_weights_block2
class Encoder(tf.keras.layers.Layer):
def __init__(self, num_layers, d_model, num_heads, dff, input_vocab_size,
rate=0.1):
super(Encoder, self).__init__()
self.d_model = d_model
self.num_layers = num_layers
pos_encoding = positional_encoding(input_vocab_size, self.d_model)
self.pos_encoding = tf.reshape(pos_encoding, [-1, self.d_model])
self.enc_layers = [EncoderLayer(d_model, num_heads, dff, rate) for _ in range(num_layers)]
self.dropout = tf.nn.dropout(input_ , 1-rate)
def call(self, x, training, mask):
seq_len = tf.shape(x)[0]
#sess.run(seq_len)
# adding embedding and position encoding.
x *= tf.sqrt(tf.cast(self.d_model, tf.float32))
x += self.pos_encoding[ :seq_len, :]
x = self.dropout(x, training=training)
for i in range(self.num_layers):
x = self.enc_layers[i](x, training, mask)
return x # ( input_seq_len, d_model)
class Decoder(tf.keras.layers.Layer):
def __init__(self, num_layers, d_model, num_heads, dff, target_vocab_size,
rate=0.1):
super(Decoder, self).__init__()
self.d_model = d_model
self.num_layers = num_layers
pos_encoding = positional_encoding(target_vocab_size, self.d_model)
self.pos_encoding = tf.reshape(pos_encoding, [-1, self.d_model])
self.dec_layers = [DecoderLayer(d_model, num_heads, dff, rate)
for _ in range(num_layers)]
self.dropout = tf.nn.dropout(input_ , 1-rate)
def call(self, x, enc_output, training, look_ahead_mask, padding_mask):
seq_len = tf.shape(x)[0]
attention_weights = {}
x *= tf.sqrt(tf.cast(self.d_model, tf.float32))
x += self.pos_encoding[ :seq_len, :]
x = self.dropout(x, training=training)
for i in range(self.num_layers):
x, block1, block2 = self.dec_layers[i](x, enc_output, training,
look_ahead_mask, padding_mask)
attention_weights['decoder_layer{}_block1'.format(i+1)] = block1
attention_weights['decoder_layer{}_block2'.format(i+1)] = block2
# x.shape == ( target_seq_len, d_model)
return x, attention_weights
class Transformer(tf.keras.Model):
def __init__(self, num_layers, d_model, num_heads, dff, input_vocab_size,
target_vocab_size, rate=0.1):
super(Transformer, self).__init__()
self.encoder = Encoder(num_layers, d_model, num_heads, dff,
input_vocab_size, rate)
self.decoder = Decoder(num_layers, d_model, num_heads, dff,
target_vocab_size, rate)
self.final_layer = tf.keras.layers.Dense(d_model)
def call(self, inp, tar, training, enc_padding_mask,
look_ahead_mask, dec_padding_mask):
enc_output = self.encoder(inp, training, enc_padding_mask) # (batch_size, inp_seq_len, d_model)
# dec_output.shape == (batch_size, tar_seq_len, d_model)
dec_output, attention_weights = self.decoder(
tar, enc_output, training, look_ahead_mask, dec_padding_mask)
final_output = self.final_layer(dec_output) # (batch_size, tar_seq_len, d_model)
return final_output, attention_weights