-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_NN.py
198 lines (177 loc) · 5.79 KB
/
test_NN.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
"""
Basically all the test cases below have same structure.
Aim : To test the forward propagation of the Neural Network
Expected : The calculations are done for a Neural Network with one hidden layer whose weights are set to 1.
Obtained : The output of the Neural Network class
Remarks: Each test case employs different set of inputs.
"""
import autodiff as ad
import numpy as np
from NN_architecture import *
import pytest
def Sigmoid(x):
return 1/(1+np.exp(-x))
def test_sigmoid_is_true():
x = 0
y=0.5
yf = Sigmoid(x)
flag = np.array_equal(y,yf)
assert flag == True
"""
def test_NN1_forward_prop_ones_init_is_true1():
model = NeuralNetLSTM(2,0,2,1)
wts = model.get_weights()
model.set_weights([np.ones_like(i()) for i in wts])
X=np.array([[0.5,0.5]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
Wf = np.ones((2,1))
Bf = np.ones(1)
output = np.dot(Snew,Wf) + Bf
print(output)
print(model.output(X)())
flag = np.array_equal(output,model.output(X)() )
assert flag==True
"""
def test_NN1_forward_prop_ones_init_is_true2():
model = NeuralNetLSTM(2,0,2,1)
wts = model.get_weights()
model.set_weights([np.ones_like(i()) for i in wts])
X=np.array([[0,0]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
Wf = np.ones((2,1))
Bf = np.ones(1)
output = np.dot(Snew,Wf) + Bf
flag = np.array_equal(output,model.output(X)() )
assert flag==True
def test_NN1_forward_prop_ones_init_is_true3():
model = NeuralNetLSTM(2,0,2,1)
wts = model.get_weights()
model.set_weights([np.ones_like(i()) for i in wts])
X=np.array([[50,50]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
Wf = np.ones((2,1))
Bf = np.ones(1)
output = np.dot(Snew,Wf) + Bf
print(output)
print(model.output(X)())
flag = np.array_equal(output,model.output(X)() )
assert flag==True
def test_NN1_forward_prop_ones_init_is_true4():
model = NeuralNetLSTM(2,0,2,1)
wts = model.get_weights()
model.set_weights([np.ones_like(i()) for i in wts])
X=np.array([[-100.1894,-100.54]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
Wf = np.ones((2,1))
Bf = np.ones(1)
output = np.dot(Snew,Wf) + Bf
flag = np.array_equal(output,model.output(X)() )
assert flag==True
def test_layer_NN1_ones_init_is_true1():
X=np.array([[100,100]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
print(Snew)
layer = lstm_layer(2,2)
wts = layer.get_weights_layer()
layer.set_params_layer([np.ones_like(i()) for i in wts])
S = ad.Variable(S,"S")
X=ad.Variable(X,"X")
output = layer.output_layer(S,X)
print(output())
flag = np.array_equal(output(),Snew)
assert flag == True
def test_layer_NN1_ones_init_is_true2():
X=np.array([[-0.1548484,0.494]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
print(Snew)
layer = lstm_layer(2,2)
wts = layer.get_weights_layer()
layer.set_params_layer([np.ones_like(i()) for i in wts])
S = ad.Variable(S,"S")
X=ad.Variable(X,"X")
output = layer.output_layer(S,X)
print(output())
flag = np.array_equal(output(),Snew)
assert flag == True
def test_layer_NN1_ones_init_is_true3():
X=np.array([[10000,1000000]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
print(Snew)
layer = lstm_layer(2,2)
wts = layer.get_weights_layer()
layer.set_params_layer([np.ones_like(i()) for i in wts])
S = ad.Variable(S,"S")
X=ad.Variable(X,"X")
output = layer.output_layer(S,X)
print(output())
flag = np.array_equal(output(),Snew)
assert flag == True
def test_layer_NN1_ones_init_is_true4():
X=np.array([[616464516,45461616]])
W = np.ones((2,2))
B = np.ones((1,2))
S = Sigmoid(np.dot(X,W) + B )
Z= Sigmoid(np.dot(X,W)+np.dot(S,W) + B)
G = np.copy(Z)
R = np.copy(G)
H = Sigmoid(np.dot(X,W)+np.dot(S*R,W)+B)
Snew = ((np.ones_like(G)-G)*H) + (Z*S)
print(Snew)
layer = lstm_layer(2,2)
wts = layer.get_weights_layer()
layer.set_params_layer([np.ones_like(i()) for i in wts])
S = ad.Variable(S,"S")
X=ad.Variable(X,"X")
output = layer.output_layer(S,X)
print(output())
flag = np.array_equal(output(),Snew)
assert flag == True