-
Notifications
You must be signed in to change notification settings - Fork 1
/
generator.py
125 lines (88 loc) · 3.73 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
import numpy as np
from keras.layers import Input,Dense
from keras.models import Model
# A toy data creation
def create_toy_folder():
xx_train =np.random.rand(100,32)
yy_train = np.random.randint(2,size=100)
for j in range(100):
np.save('folder1\\aa_x_'+str(j)+'.npy',xx_train[j,:])
np.save('folder2\\aa_y_' + str(j) + '.npy', yy_train[j])
xx_test = np.random.rand(50, 32)
yy_test = np.random.randint(2,size=50)
print (yy_test)
for j in range(50):
np.save('foldertest\\aa_x_'+str(j)+'.npy',xx_test[j,:])
np.save('foldertest\\aa_y_' + str(j) + '.npy', yy_test[j])
return xx_train,yy_train,xx_test,yy_test
class organizer():
def __init__(self,folder, amount_files, batch_size,feat_dim):
self.mode_folder =folder
self.y_folder = folder+"_y"
self.nfiles =amount_files
self.batch_size = batch_size
self.feat_dim =feat_dim
self.xtrain = np.empty((0, self.feat_dim))
self.ytrain = []
def __len__(self):
return len(self.list_IDs)
def get_data_item(self, index):
x = np.expand_dims(np.load(self.mode_folder +"\\aa_x_"+str(index) + '.npy'),axis=0) #expand_dims only if needed
y = np.load(self.y_folder + "\\aa_y_" + str(index) + '.npy').item() # reading ndarray as scalar since np.load brings only arrays
return x,y
def data_genereator(self):
counter = 0
while True:
for index in range(self.nfiles):
if self.xtrain.shape[0]==self.batch_size:
self.xtrain = np.empty((0, self.feat_dim))
self.ytrain= []
x,y =self.get_data_item(index)
self.xtrain = np.concatenate(( self.xtrain, x), axis=0)
self.ytrain.append(y)
counter = counter + 1
if (counter % self.batch_size == 0):
x= self.xtrain
y= self.ytrain
yield x, np.array(y)
def get_data_test(self, index):
x = np.expand_dims(np.load(self.mode_folder + "\\aa_x_" + str(index) + '.npy'),
axis=0) # expand_dims only if needed
return x
def data_genereator_test(self):
counter = 0
while True:
for index in range(self.nfiles):
if self.xtrain.shape[0]==self.batch_size:
self.xtrain = np.empty((0, self.feat_dim))
x =self.get_data_test(index)
self.xtrain = np.concatenate(( self.xtrain, x), axis=0)
counter = counter + 1
if (counter % self.batch_size == 0):
x= self.xtrain
yield x
if __name__ == '__main__':
# The way I created the toy data
# x_train,ytr,xte,yte =create_toy_folder()
train_p ="your train data x (y if needed"
test_p ="your test data x"
ntest = 50
batch_size =8
nitems=100
feat_dim=32
organ_train = organizer(train_p, nitems, 8, feat_dim)
organ_test = organizer(test_p, ntest, 8, feat_dim )
gen =organ_train.data_genereator()
inp = Input(shape=( 32,))
x =Dense(5,activation="relu")(inp)
x = Dense(1, activation='sigmoid')(x)
model = Model(inputs=[inp], outputs=[x])
model.compile(loss='binary_crossentropy', optimizer="RMSprop", metrics=['accuracy'])
print (model.summary())
#Train
model.fit_generator(gen, steps_per_epoch=nitems // batch_size, verbose=1)
#Test
gent = organ_test.data_genereator_test()
score = model.predict_generator(gent, ntest // batch_size, verbose=1)
print (score)
print ("haha")