Malaria/train.py at master · We-Gold/Malaria · GitHub

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
import keras
from keras.preprocessing.image import ImageDataGenerator
from keras.layers import Dense,GlobalAveragePooling2D
from keras.applications import MobileNet
from keras.preprocessing import image
from keras.applications.mobilenet import preprocess_input
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Model
from keras.optimizers import Adam
from keras.callbacks import ModelCheckpoint

train_datagen = ImageDataGenerator(rescale=1./255,
                                   shear_range=0.2,
                                   zoom_range=0.2,
                                   horizontal_flip=True)

test_datagen = ImageDataGenerator(rescale=1./255)

train_generator = train_datagen.flow_from_directory('cell_images/train',
                                                    target_size=(100,100),
                                                    batch_size=32,
                                                    class_mode='binary')

validation_generator = test_datagen.flow_from_directory('cell_images/test',
                                                    target_size=(100,100),
                                                    batch_size=32,
                                                    class_mode='binary')


base_model = MobileNet(weights = 'imagenet',include_top = False)

x = base_model.output
x = GlobalAveragePooling2D()(x)
x = Dense(512,activation = 'relu')(x)
x = Dense(256,activation = 'relu')(x)
preds = Dense(1,activation = 'sigmoid')(x)

model = Model(inputs = base_model.input,outputs = preds)

for layer in model.layers[:20]:
    layer.trainable = False
for layer in model.layers[20:]:
    layer.trainable = True

model.compile(optimizer = 'Adam',loss = 'binary_crossentropy',metrics = ['accuracy'])

checkpoints = ModelCheckpoint("checkpoints/weights.{epoch:02d}.h5",
                                          save_weights_only = False,
                                          verbose = 1)

#step_size_train = train_generator.n//train_generator.batch_size


model.fit_generator(train_generator,
                        steps_per_epoch=8000,
                        epochs=5,
                        validation_data=validation_generator,
                        validation_steps=800,
                        callbacks = [checkpoints])

model.save("model.h5")