-
Notifications
You must be signed in to change notification settings - Fork 4
/
convNetKerasLarge.py
131 lines (106 loc) · 4.72 KB
/
convNetKerasLarge.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
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from sklearn.model_selection import train_test_split
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras.callbacks import TensorBoard
from keras import optimizers
from skimage import io
from skimage.transform import resize
from keras.utils import to_categorical
import numpy as np
import tensorflow as tf
import random
import glob
import coremltools
random_seed = 1
tf.set_random_seed(random_seed)
np.random.seed(random_seed)
batch_size = 32
num_classes = 2
epochs = 80
log_filepath = './deepKerasLog'
nCategorySamples = 4000
positiveSamples = glob.glob('workspace/SoBr/data/resize_frontal_face/yes/*')[0:nCategorySamples]
negativeSamples = glob.glob('workspace/SoBr/data/resize_frontal_face/no/*')[0:nCategorySamples]
nImageRows = 106
nImageCols = 106
nChannels = 3
negativeSamples = random.sample(negativeSamples, len(positiveSamples))
X_train = []
Y_train = []
for i in range(len(positiveSamples)):
X_train.append(resize(io.imread(positiveSamples[i]), (nImageRows, nImageCols)))
Y_train.append(1)
if i % 1000 == 0:
print('Reading positive image number ', i)
for i in range(len(negativeSamples)):
X_train.append(resize(io.imread(negativeSamples[i]), (nImageRows, nImageCols)))
Y_train.append(0)
if i % 1000 == 0:
print('Reading negative image number ', i)
X_train = np.array(X_train)
Y_train = np.array(Y_train)
X_train, X_test, Y_train, Y_test = train_test_split(X_train, Y_train, test_size=0.30, random_state=42)
mean = X_train.mean(axis=0).mean(axis=0).mean(axis=0)
#std = X_train.std(axis=0).mean(axis=0).mean(axis=0)
mean = np.array([0.5,0.5,0.5])
std = np.array([1,1,1])
X_train = X_train.astype('float')
X_test = X_test.astype('float')
for i in range(3):
X_train[:,:,:,i] = (X_train[:,:,:,i]- mean[i]) / std[i]
X_test[:,:,:,i] = (X_test[:,:,:,i]- mean[i]) / std[i]
num_iterations = int(len(X_train)/batch_size) + 1
Y_train = to_categorical(Y_train, num_classes)
Y_test = to_categorical(Y_test, num_classes)
modelInputShape = (nImageRows, nImageCols, nChannels)
model = Sequential()
model.add(Conv2D(8,kernel_size=(3,3), activation='relu', strides=(1,1), padding='same', input_shape=modelInputShape))
model.add(MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='valid'))
model.add(Dropout(0.25))
model.add(Conv2D(16,kernel_size=(3,3), padding='same', activation='relu'))
model.add(MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='valid'))
model.add(Dropout(0.25))
model.add(Conv2D(16,kernel_size=(3,3), padding='same', activation='relu'))
model.add(MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='valid'))
model.add(Conv2D(8,kernel_size=(3,3), padding='same', activation='relu'))
model.add(MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='valid'))
#model.add(Conv2D(8,kernel_size=(3,3), padding='same', activation='relu'))
#model.add(MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='valid'))
model.add(Flatten())
model.add(Dense(10, activation='relu'))
model.add(Dense(2, activation='softmax'))
sgd = optimizers.SGD(lr=.001, momentum=0.9, decay=0.000005, nesterov=False)
model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])
print(model.summary())
#datagen = ImageDataGenerator(
# featurewise_center=True,
# featurewise_std_normalization=True,
# rotation_range=0,
# width_shift_range=0.05,
# height_shift_range=0.07,
# zoom_range=0.05,
# horizontal_flip=True)
#datagen.fit(X_train)
tensorBoardCallback = TensorBoard(log_dir=log_filepath, histogram_freq=0)
callbacks = [tensorBoardCallback]
model.fit(X_train, Y_train, batch_size=batch_size, epochs=epochs, validation_data=(X_test,Y_test))
#model.fit_generator(datagen.flow(X_train, Y_train,batch_size=batch_size),
# steps_per_epoch=1*num_iterations,
# epochs=epochs,
# callbacks=callbacks,
# validation_data=(X_test, Y_test))
score=model.evaluate(X_test, Y_test, verbose=0)
print('Test loss: ', score[0])
print('Test accuracy: ', score[1])
coreml_model = coremltools.converters.keras.convert(model,
input_names=['image'],
output_names=['output'],
class_labels=['Negative', 'Drunk'],
image_input_names='image',
image_scale=1/255.0,
red_bias=-0.5,
green_bias=-0.5,
blue_bias=-0.5)
coreml_model.save('DrunkKerasModel.mlmodel')