Skip to content

Commit

Permalink
commit #34: refactoring resnet
Browse files Browse the repository at this point in the history
  • Loading branch information
@BIGBALLON
BIGBALLON committed Apr 22, 2018
1 parent 455cc41 commit 7c644e0
Showing 1 changed file with 73 additions and 81 deletions.
154 changes: 73 additions & 81 deletions 4_Residual_Network/ResNet_keras.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,111 +4,100 @@
from keras.preprocessing.image import ImageDataGenerator
from keras.layers.normalization import BatchNormalization
from keras.layers import Conv2D, Dense, Input, add, Activation, GlobalAveragePooling2D
from keras.initializers import he_normal
from keras.callbacks import LearningRateScheduler, TensorBoard
from keras.callbacks import LearningRateScheduler, TensorBoard, ModelCheckpoint
from keras.models import Model
from keras import optimizers
from keras import regularizers
from keras import optimizers, regularizers

stack_n = 5
num_classes = 10
img_rows, img_cols = 32, 32
img_channels = 3
stack_num = 18
batch_size = 128
epochs = 200
iterations = 391
weight_decay = 0.0005
log_filepath = r'./resnet50/'
iterations = 50000 // batch_size
weight_decay = 0.0001
mean = [125.307, 122.95, 113.865]
std = [62.9932, 62.0887, 66.7048]

def scheduler(epoch):
if epoch <= 60:
return 0.05
if epoch <= 120:
if epoch < 80:
return 0.1
if epoch < 150:
return 0.01
if epoch <= 160:
return 0.002
return 0.0004
return 0.001

def color_preprocessing(x_train,x_test):
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
mean = [125.307, 122.95, 113.865]
std = [62.9932, 62.0887, 66.7048]
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]

return x_train, x_test

def residual_network(img_input,classes_num=10,stack_n=18):
def residual_block(x,shape,increase_filter=False):
output_filter_num = shape[1]
if increase_filter:
first_stride = (2,2)
def residual_network(img_input,classes_num=10,stack_n=5):
def residual_block(intput,out_channel,increase=False):
if increase:
stride = (2,2)
else:
first_stride = (1,1)
stride = (1,1)

pre_bn = BatchNormalization()(x)
pre_bn = BatchNormalization()(intput)
pre_relu = Activation('relu')(pre_bn)

conv_1 = Conv2D(output_filter_num,
kernel_size=(3,3),
strides=first_stride,
padding='same',
kernel_initializer=he_normal(),
kernel_regularizer=regularizers.l2(weight_decay)
)(pre_relu)
conv_1 = Conv2D(out_channel,kernel_size=(3,3),strides=stride,padding='same',
kernel_initializer="he_normal",
kernel_regularizer=regularizers.l2(weight_decay))(pre_relu)
bn_1 = BatchNormalization()(conv_1)
relu1 = Activation('relu')(bn_1)
conv_2 = Conv2D(output_filter_num,
kernel_size=(3,3),
strides=(1,1),
padding='same',
kernel_initializer=he_normal(),
kernel_regularizer=regularizers.l2(weight_decay)
)(relu1)
if increase_filter:
projection = Conv2D(output_filter_num,
kernel_size=(1,1),
strides=(2,2),
padding='same',
kernel_initializer=he_normal(),
kernel_regularizer=regularizers.l2(weight_decay)
)(x)
conv_2 = Conv2D(out_channel,kernel_size=(3,3),strides=(1,1),padding='same',
kernel_initializer="he_normal",
kernel_regularizer=regularizers.l2(weight_decay))(relu1)
if increase:
projection = Conv2D(out_channel,
kernel_size=(1,1),
strides=(2,2),
padding='same',
kernel_initializer="he_normal",
kernel_regularizer=regularizers.l2(weight_decay))(intput)
block = add([conv_2, projection])
else:
block = add([conv_2,x])
block = add([intput,conv_2])
return block

x = Conv2D(filters=16,
kernel_size=(3,3),
strides=(1,1),
padding='same',
kernel_initializer=he_normal(),
kernel_regularizer=regularizers.l2(weight_decay)
)(img_input)
# build model
# total layers = stack_n * 3 * 2 + 2
# stack_n = 5 by default, total layers = 32
# input: 32x32x3 output: 32x32x16
x = Conv2D(filters=16,kernel_size=(3,3),strides=(1,1),padding='same',
kernel_initializer="he_normal",
kernel_regularizer=regularizers.l2(weight_decay))(img_input)

for _ in range(0,stack_n):
x = residual_block(x,[16,16])
# input: 32x32x16 output: 32x32x16
for _ in range(stack_n):
x = residual_block(x,16,False)

x = residual_block(x,[16,32],increase_filter=True)
# input: 32x32x16 output: 16x16x32
x = residual_block(x,32,True)
for _ in range(1,stack_n):
x = residual_block(x,[16,32])

x = residual_block(x,[32,64],increase_filter=True)
x = residual_block(x,32,False)

# input: 16x16x32 output: 8x8x64
x = residual_block(x,64,True)
for _ in range(1,stack_n):
x = residual_block(x,[32,64])
x = residual_block(x,64,False)

x = BatchNormalization()(x)
x = Activation('relu')(x)
x = GlobalAveragePooling2D()(x)
x = Dense(classes_num,activation='softmax',kernel_initializer=he_normal(),kernel_regularizer=regularizers.l2(weight_decay))(x)

# input: 64 output: 10
x = Dense(classes_num,activation='softmax',
kernel_initializer="he_normal",
kernel_regularizer=regularizers.l2(weight_decay))(x)
return x


def color_preprocessing(x_train,x_test):
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
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]
return x_train, x_test

if __name__ == '__main__':

# load data
(x_train, y_train), (x_test, y_test) = cifar10.load_data()
y_train = keras.utils.to_categorical(y_train, num_classes)
Expand All @@ -119,29 +108,32 @@ def residual_block(x,shape,increase_filter=False):

# build network
img_input = Input(shape=(img_rows,img_cols,img_channels))
output = residual_network(img_input,num_classes,18)
resnet = Model(img_input, output)
output = residual_network(img_input,num_classes,stack_n)
resnet = Model(img_input, output)
print(resnet.summary())

# set optimizer
sgd = optimizers.SGD(lr=.1, momentum=0.9, nesterov=True)
resnet.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])

# set callback
tb_cb = TensorBoard(log_dir=log_filepath, histogram_freq=0)
change_lr = LearningRateScheduler(scheduler)
cbks = [change_lr,tb_cb]
cbks = [TensorBoard(log_dir='./resnet_32/', histogram_freq=0),
LearningRateScheduler(scheduler),
ModelCheckpoint('./checkpoint-{epoch}.h5', save_best_only=False, mode='auto', period=10)]

# set data augmentation
print('Using real-time data augmentation.')
datagen = ImageDataGenerator(horizontal_flip=True,
width_shift_range=0.125,height_shift_range=0.125,fill_mode='constant',cval=0.)
width_shift_range=0.125,
height_shift_range=0.125,
fill_mode='constant',cval=0.)

datagen.fit(x_train)

# start training
resnet.fit_generator(datagen.flow(x_train, y_train,batch_size=batch_size),
steps_per_epoch=iterations,
epochs=epochs,
callbacks=cbks,
validation_data=(x_test, y_test))
steps_per_epoch=iterations,
epochs=epochs,
callbacks=cbks,
validation_data=(x_test, y_test))
resnet.save('resnet.h5')

0 comments on commit 7c644e0

Please sign in to comment.