forked from CSAILVision/places365
-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathfinetune_vgg16.py
146 lines (126 loc) · 4.84 KB
/
finetune_vgg16.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
import tensorflow as tf
from keras.models import Sequential
from keras.utils import np_utils
from keras.layers.core import Flatten, Dense, Dropout
from keras.layers.convolutional import Convolution2D, MaxPooling2D, ZeroPadding2D
from keras.optimizers import SGD
from keras.callbacks import TensorBoard
import cv2, numpy as np
import pickle
import sys
import random
import os
# Don't hog GPU
config = tf.ConfigProto()
config.gpu_options.allow_growth=True
sess = tf.Session(config=config)
keras.backend.set_session(sess)
# Source for vgg16 implementation : https://gist.github.com/baraldilorenzo/07d7802847aaad0a35d3
def ready_model(weights_path, n_classes):
model = Sequential()
model.add(ZeroPadding2D((1,1),input_shape=(3,224,224)))
model.add(Convolution2D(64, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(64, 3, 3, activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(128, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(128, 3, 3, activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(512, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(512, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(512, 3, 3, activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(512, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(512, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(512, 3, 3, activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Flatten())
model.add(Dense(4096, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(4096, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(365, activation='softmax'))
model.load_weights(weights_path)
# Freeze all layers except the last two
model.pop()
model.pop()
model.add(Dropout(0.5))
model.add(Dense(n_classes, activation='softmax'))
sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(optimizer=sgd, loss='categorical_crossentropy',metrics=['accuracy'])
return model
def process_image(img_path):
im = cv2.resize(cv2.imread(img_path), (224, 224)).astype(np.float32)
im[:,:,0] -= 103.939
im[:,:,1] -= 116.779
im[:,:,2] -= 123.68
im = im.transpose((2,0,1))
return im
def train_test_set(image_filenames, n_classes, ratio = 0.8):
X = []
y = []
random.shuffle(image_filenames.keys())
for img in image_filenames.keys():
X.append(process_image(img))
y.append(image_filenames[img])
split_ratio = int(len(X) * ratio)
X_train, y_train = X[:split_ratio], y[:split_ratio]
X_test, y_test = X[split_ratio:], y[split_ratio:]
# to_categorical
X_train, y_train = np.array(X_train), np_utils.to_categorical(np.array(y_train), n_classes)
X_test, y_test = np.array(X_test), np_utils.to_categorical(np.array(y_test), n_classes)
return (X_train, y_train), (X_test, y_test)
def ready_data(base_dir):
# Link class names with labels
name_index_mapping = {}
index_name_mapping = {}
file_name_mapping = {}
labels = []
for root, subdirs, files in os.walk(base_dir):
labels = subdirs[:]
break
for class_index, class_name in enumerate(labels):
name_index_mapping[class_name] = class_index
index_name_mapping[class_index] = class_name
# Walk through directory, gathering names of images with their labels
n_classes = len(name_index_mapping)
for root, subdirs, files in os.walk(base_dir):
for filename in files:
file_path = os.path.join(root, filename)
assert file_path.startswith(base_dir)
suffix = file_path[len(base_dir):]
suffix = suffix.lstrip("/")
label = suffix.split("/")[0]
file_name_mapping[file_path] = name_index_mapping[label]
return file_name_mapping, n_classes
def fine_tune(model, n_classes, file_name_map):
n_eopchs = 10
batch_size = 32
(X_train, y_train), (X_test, y_test) = train_test_set(file_name_map, n_classes)
model.fit(X_train,y_train,
nb_epoch = n_eopchs,
batch_size = batch_size,
callbacks = [TensorBoard(log_dir='/tmp/places_vgg16_finetune')])
loss_and_metrics = model.evaluate(X_test, y_test, batch_size = 32)
model.save("finetuned_keras.h5")
return loss_and_metrics
if __name__ == "__main__":
file_name_map, n_classes = ready_data(sys.argv[1])
model = ready_model('models/places/places_vgg_keras.h5', n_classes)
acc = fine_tune(model, n_classes, file_name_map)
print "Accuracy was ", acc