-
Notifications
You must be signed in to change notification settings - Fork 1
/
Recurrent_Unit.py
201 lines (179 loc) · 6.43 KB
/
Recurrent_Unit.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
import numpy as np
import theano
import theano.tensor as T
from util import init_weight
class GRU:
def __init__(self, Mi, Mo, activation, state=None):
self.Mi = Mi
self.Mo = Mo
self.f = activation
self.__setstate__(state)
def __getstate__(self):
return tuple([p.get_value() for p in self.params])
def __setstate__(self, state=None):
if state:
Wxr, Whr, br, Wxz, Whz, bz, Wxh, Whh, bh, h0 = state
else:
Wxr, Whr, br, Wxz, Whz, bz, Wxh, Whh, bh, h0 = init_weight(self.Mi, self.Mo), init_weight(self.Mo, self.Mo), np.zeros(self.Mo), init_weight(self.Mi, self.Mo), init_weight(self.Mo, self.Mo), np.zeros(self.Mo), init_weight(self.Mi, self.Mo), init_weight(self.Mo, self.Mo), np.zeros(self.Mo), np.zeros(self.Mo)
self.Wxr = theano.shared(Wxr)
self.Whr = theano.shared(Whr)
self.br = theano.shared(br)
self.Wxz = theano.shared(Wxz)
self.Whz = theano.shared(Whz)
self.bz = theano.shared(bz)
self.Wxh = theano.shared(Wxh)
self.Whh = theano.shared(Whh)
self.bh = theano.shared(bh)
self.h0 = theano.shared(h0)
self.params = [self.Wxr, self.Whr, self.br, self.Wxz, self.Whz, self.bz, self.Wxh, self.Whh, self.bh, self.h0]
def get_ht(self, xWxr_t, xWxz_t, xWxh_t, h_t1):
r = T.nnet.sigmoid(xWxr_t + h_t1.dot(self.Whr) + self.br)
z = T.nnet.sigmoid(xWxz_t + h_t1.dot(self.Whz) + self.bz)
hhat = self.f(xWxh_t + (r * h_t1).dot(self.Whh) + self.bh)
h = (1 - z) * h_t1 + z * hhat
return h
def recurrence(self, xWxr_t, xWxz_t, xWxh_t, is_start, h_t1, h0):
h_t = T.switch(
T.eq(is_start, 1),
self.get_ht(xWxr_t, xWxz_t, xWxh_t, h0),
self.get_ht(xWxr_t, xWxz_t, xWxh_t, h_t1)
)
return h_t
def output(self, Xflat, startPoints):
XWxr = Xflat.dot(self.Wxr)
XWxz = Xflat.dot(self.Wxz)
XWxh = Xflat.dot(self.Wxh)
h, _ = theano.scan(
fn=self.recurrence,
sequences=[XWxr, XWxz, XWxh, startPoints],
outputs_info=[self.h0],
non_sequences=[self.h0],
n_steps=Xflat.shape[0],
)
return h
class LSTM:
def __init__(self, Mi, Mo, activation):
self.Mi = Mi
self.Mo = Mo
self.f = activation
# numpy init
Wxi = init_weight(Mi, Mo)
Whi = init_weight(Mo, Mo)
Wci = init_weight(Mo, Mo)
bi = np.zeros(Mo)
Wxf = init_weight(Mi, Mo)
Whf = init_weight(Mo, Mo)
Wcf = init_weight(Mo, Mo)
bf = np.zeros(Mo)
Wxc = init_weight(Mi, Mo)
Whc = init_weight(Mo, Mo)
bc = np.zeros(Mo)
Wxo = init_weight(Mi, Mo)
Who = init_weight(Mo, Mo)
Wco = init_weight(Mo, Mo)
bo = np.zeros(Mo)
c0 = np.zeros(Mo)
h0 = np.zeros(Mo)
# theano vars
self.Wxi = theano.shared(Wxi)
self.Whi = theano.shared(Whi)
self.Wci = theano.shared(Wci)
self.bi = theano.shared(bi)
self.Wxf = theano.shared(Wxf)
self.Whf = theano.shared(Whf)
self.Wcf = theano.shared(Wcf)
self.bf = theano.shared(bf)
self.Wxc = theano.shared(Wxc)
self.Whc = theano.shared(Whc)
self.bc = theano.shared(bc)
self.Wxo = theano.shared(Wxo)
self.Who = theano.shared(Who)
self.Wco = theano.shared(Wco)
self.bo = theano.shared(bo)
self.c0 = theano.shared(c0)
self.h0 = theano.shared(h0)
self.params = [
self.Wxi,
self.Whi,
self.Wci,
self.bi,
self.Wxf,
self.Whf,
self.Wcf,
self.bf,
self.Wxc,
self.Whc,
self.bc,
self.Wxo,
self.Who,
self.Wco,
self.bo,
self.c0,
self.h0,
]
def get_ht_ct(self, xWxi_t, xWxf_t, xWxc_t, xWxo_t, h_t1, c_t1):
i_t = T.nnet.sigmoid(xWxi_t + h_t1.dot(self.Whi) + c_t1.dot(self.Wci) + self.bi)
f_t = T.nnet.sigmoid(xWxf_t + h_t1.dot(self.Whf) + c_t1.dot(self.Wcf) + self.bf)
c_t = f_t * c_t1 + i_t * T.tanh(xWxc_t + h_t1.dot(self.Whc) + self.bc)
o_t = T.nnet.sigmoid(xWxo_t + h_t1.dot(self.Who) + c_t.dot(self.Wco) + self.bo)
h_t = o_t * T.tanh(c_t)
return h_t, c_t
def recurrence(self, xWxi_t, xWxf_t, xWxc_t, xWxo_t, is_start, h_t1, c_t1, h0, c0):
h_t_c_t = T.switch(
T.eq(is_start, 1),
self.get_ht_ct(xWxi_t, xWxf_t, xWxc_t, xWxo_t, h0, c0),
self.get_ht_ct(xWxi_t, xWxf_t, xWxc_t, xWxo_t, h_t1, c_t1)
)
return h_t_c_t[0], h_t_c_t[1]
def output(self, Xflat, startPoints):
# Xflat should be (NT, D)
# calculate X after multiplying input weights
XWxi = Xflat.dot(self.Wxi)
XWxf = Xflat.dot(self.Wxf)
XWxc = Xflat.dot(self.Wxc)
XWxo = Xflat.dot(self.Wxo)
[h, c], _ = theano.scan(
fn=self.recurrence,
sequences=[XWxi, XWxf, XWxc, XWxo, startPoints],
outputs_info=[self.h0, self.c0],
non_sequences=[self.h0, self.c0],
n_steps=Xflat.shape[0],
)
return h
class SimpleRecurrentLayer:
def __init__(self, Mi, Mo, activation):
self.Mi = Mi
self.Mo = Mo
self.f = activation
# numpy init
Wxh = init_weight(Mi, Mo)
Whh = init_weight(Mo, Mo)
b = np.zeros(Mo)
h0 = np.zeros(Mo)
# theano vars
self.Wxh = theano.shared(Wxh)
self.Whh = theano.shared(Whh)
self.b = theano.shared(b)
self.h0 = theano.shared(h0)
self.params = [self.Wxh, self.Whh, self.b, self.h0]
def get_ht(self, xWxh_t, h_t1):
return self.f(xWxh_t + h_t1.dot(self.Whh) + self.b)
def recurrence(self, xWxh_t, is_start, h_t1, h0):
h_t = T.switch(
T.eq(is_start, 1),
self.get_ht(xWxh_t, h0),
self.get_ht(xWxh_t, h_t1)
)
return h_t
def output(self, Xflat, startPoints):
# Xflat should be (NT, D)
# calculate X after multiplying input weights
XWxh = Xflat.dot(self.Wxh)
h, _ = theano.scan(
fn=self.recurrence,
sequences=[XWxh, startPoints],
outputs_info=[self.h0],
non_sequences=[self.h0],
n_steps=Xflat.shape[0],
)
return h