-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtrain.py
177 lines (144 loc) · 6.78 KB
/
train.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
import time
import os.path
import sys
import pathlib
import numpy as np
from keras.callbacks import TensorBoard, ModelCheckpoint, EarlyStopping, CSVLogger
from sklearn.metrics import confusion_matrix
from sklearn.model_selection import KFold
from model import Model
from data import Data
def train(model_name, num_frames=48, num_features=4, saved_model=None,
image_shape=None, num_samples=70, save_trained_model=True, fold_validate=False,
load_to_memory=False, batch_size=1, nb_epoch=100, drop_out=0.3):
# Helper: TensorBoard
tb = TensorBoard(log_dir=os.path.join('data', 'logs', model_name))
# Helper: Stop when we stop learning.
early_stopper = EarlyStopping(patience=5)
# Helper: Save results.
timestamp = time.time()
csv_logger = CSVLogger(os.path.join('data', 'logs', model_name + '-' + 'training-' +
str(timestamp) + '.log'))
# Get the data and process it.
data = Data(num_frames=num_frames, image_shape=image_shape)
rm = Model(model_name, num_frames=num_frames,
saved_model=saved_model, image_shape=image_shape)
if fold_validate:
if model_name == 'lstm':
X, y = data.load_extracted_data(split=False)
kf = KFold(n_splits=10)
tn, fp, fn, tp = 0, 0, 0, 0
count = 1
for train, test in kf.split(X, y):
rm.model.fit(
X[train],
y[train],
batch_size=batch_size,
validation_data=(X[test], y[test]),
verbose=0,
callbacks=[tb, early_stopper, csv_logger],
epochs=nb_epoch)
test_loss, test_acc = rm.model.evaluate(X[test], y[test])
print(
'Fold {}\n---- Test Accuracy: {}\n---- Test Loss: {}\n'.format(count, test_acc, test_loss))
y_pred = rm.model.predict(X[test])
y_pred = (y_pred > 0.5)
# tn, fp, fn, tp
tn1, fp1, fn1, tp1 = confusion_matrix(
y[test], y_pred, labels=[0, 1]).ravel()
tn += tn1
fp += fp1
fn += fn1
tp += tp1
count += 1
print('tn: {}, fp: {}, fn: {}, tp: {}'.format(tn, fp, fn, tp))
else:
if model_name == 'lstm':
X_train, X_test, y_train, y_test = data.load_extracted_data()
rm.model.fit(
X_train,
y_train,
batch_size=batch_size,
validation_data=(X_test, y_test),
verbose=1,
callbacks=[tb, early_stopper, csv_logger],
epochs=nb_epoch)
test_loss, test_acc = rm.model.evaluate(X_test, y_test)
print('Test Accuracy: {}\nTest Loss: {}'.format(test_acc, test_loss))
y_pred = rm.model.predict(X_test)
y_pred = (y_pred > 0.5)
# tn, fp, fn, tp
cm = confusion_matrix(y_test, y_pred, labels=[0, 1])
print(cm)
if save_trained_model:
save_model_name = 'model-{}-{}.h5'.format(model_name, test_acc)
if not os.path.isdir(os.path.join('data', 'trained')):
pathlib.Path(os.path.join('data', 'trained')).mkdir(
parents=True, exist_ok=True)
if not os.path.isfile(os.path.join('data', 'trained', save_model_name)):
rm.model.save(os.path.join('data', 'trained', save_model_name))
elif model_name == 'lrcn':
data.load_image_train_split()
steps_per_epoch = (data.num_samples * 0.7) // batch_size
validation_steps = (data.num_samples * 0.3) // batch_size
generator = data.frame_generator(batch_size, 'train')
val_generator = data.frame_generator(batch_size, 'test')
rm.model.fit_generator(
generator=generator,
steps_per_epoch=steps_per_epoch,
epochs=nb_epoch,
verbose=1,
callbacks=[tb, early_stopper, csv_logger],
validation_data=val_generator,
validation_steps=validation_steps,
workers=4)
if save_trained_model:
save_model_name = 'model-{}.h5'.format(model_name)
if not os.path.isdir(os.path.join('data', 'trained')):
pathlib.Path(os.path.join('data', 'trained')).mkdir(
parents=True, exist_ok=True)
if not os.path.isfile(os.path.join('data', 'trained', save_model_name)):
rm.model.save(os.path.join('data', 'trained', save_model_name))
test_generator = data.frame_generator(batch_size, 'test')
y_pred = []
y_test = []
for test_sample, label in test_generator:
y_pred += [np.around(rm.model.predict(test_sample))]
y_test += [label]
# tn, fp, fn, tp
tn.fp,fn,tp = confusion_matrix(y_test, y_pred, labels=[0, 1]).ravel()
print('tn: {}, fp: {}, fn: {}, tp: {}'.format(tn, fp, fn, tp))
def test(model_name, num_frames=48, num_features=4, saved_model=None, image_shape=None, num_samples=70, save_trained_model=True, fold_validate=False,load_to_memory=False, batch_size=1, nb_epoch=100, drop_out=0.3):
rm = Model(model_name, num_frames=num_frames, saved_model=saved_model, image_shape=image_shape)
# Get the data and process it.
data = Data(num_frames=num_frames, image_shape=image_shape)
data.load_image_train_split()
test_generator = data.frame_generator(batch_size, 'test')
y_pred = []
y_test = []
for test_sample, label in test_generator:
res = np.around(rm.model.predict(test_sample))
y_test.append(label)
y_pred.append(res)
print('true: {}, predict: {}'.format(label, res))
# tn, fp, fn, tp
tn,fp,fn,tp = confusion_matrix(y_test, y_pred, labels=[0, 1]).ravel()
print('tn: {}, fp: {}, fn: {}, tp: {}'.format(tn, fp, fn, tp))
def main():
model_name = 'lrcn'
saved_model = os.path.join('data', 'trained','model-lrcn.h5') # None or weights file
save_trained_model = True
batch_size = 5
nb_epoch = 100
num_frames = 120
image_shape = (160, 120, 3)
fold_validate = False
num_features = 4
if len(sys.argv) > 1:
model_name = sys.argv[1]
# train(model_name, num_frames=num_frames, saved_model=saved_model, image_shape=image_shape, fold_validate=fold_validate,
# batch_size=batch_size, nb_epoch=nb_epoch, num_features=num_features, save_trained_model=save_trained_model)
test(model_name, num_frames=num_frames, saved_model=saved_model, image_shape=image_shape, fold_validate=fold_validate,
batch_size=batch_size, nb_epoch=nb_epoch, num_features=num_features, save_trained_model=save_trained_model)
if __name__ == '__main__':
main()