-
Notifications
You must be signed in to change notification settings - Fork 2
/
cnn_model.py
84 lines (56 loc) · 2.45 KB
/
cnn_model.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
from keras.models import Sequential
from keras.layers.core import Dense, Flatten, Dropout
from keras.layers.convolutional import Conv2D
import keras
from keras.callbacks import ModelCheckpoint
from sklearn.model_selection import train_test_split
import os
import constants
import data_manager
print (keras.__version__)
# paper link for the convolutional model - https://arxiv.org/pdf/1604.07316.pdf
def _create_model():
nrows = constants.IMAGE_HEIGHT
ncols = constants.IMAGE_WIDTH
img_channels = 3 # color channels
output_size = 2
model = Sequential()
model.add(Dropout(0.35, input_shape=(nrows, ncols, img_channels)))
model.add(Conv2D(filters=24, kernel_size=(5, 5), strides=(2, 2), padding='valid', activation='elu'))
model.add(Conv2D(filters=36, kernel_size=(5, 5), strides=(2, 2), padding='valid', activation='elu'))
model.add(Conv2D(filters=48, kernel_size=(5, 5), strides=(2, 2), padding='valid', activation='elu'))
model.add(Conv2D(filters=64, kernel_size=(3, 3), strides=(1, 1), padding='valid', activation='elu'))
model.add(Conv2D(filters=64, kernel_size=(3, 3), strides=(1, 1), padding='valid', activation='elu'))
model.add(Dropout(0.35))
model.add(Flatten())
model.add(Dense(100, activation='relu'))
model.add(Dense(50, activation='relu'))
model.add(Dense(10, activation='relu'))
model.add(Dense(output_size))
model.summary()
model.compile(loss=keras.losses.mean_squared_error,
optimizer=keras.optimizers.Adam(lr=0.001),
metrics=['accuracy'])
return model
def get_model():
if (os.path.isfile(constants.FINAL_MODEL_FILEPATH)):
model = keras.models.load_model('my_model.h5')
else:
model = _create_model()
return model
def train_model(model):
"""
Training data will consist 85% of total data.
The last 15% will be used for validation
:param model: kenas cnn model
:return:
"""
X, Y = data_manager.read_data_from_file(split = True)
X_train, X_valid, Y_train, Y_valid = train_test_split(X, Y, test_size = 0.15, random_state = 1)
model = get_model()
checkpoint = ModelCheckpoint('model-{epoch:03d}.h5', monitor='val_loss', verbose=1, mode='auto')
model.fit(X_train, Y_train, epochs=8, batch_size=64, verbose=1,
validation_data=(X_valid, Y_valid), callbacks=[checkpoint])
#save the model after training
model.save(constants.FINAL_MODEL_FILEPATH)
# model.summary()