def InceptionV1(include_top=True,
#weights= None,
weights= 'imagenet',
input_tensor=None,
input_shape=None,
pooling=None,
classes=2):
"""Instantiates the Inception v1 architecture.
This architecture is defined in:
Going deeper with convolutions
Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed,
Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, Andrew Rabinovich.
http://arxiv.org/abs/1409.4842v1
Optionally loads weights pre-trained
on ImageNet. Note that when using TensorFlow,
for best performance you should set
`image_data_format="channels_last"` in your Keras config
at ~/.keras/keras.json.
The model and the weights are compatible with both
TensorFlow and Theano. The data format
convention used by the model is the one
specified in your Keras config file.
Note that the default input image size for this model is 224x224.
Arguments:
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization)
or "imagenet" (pre-training on ImageNet).
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or `(3, 224, 224)` (with `channels_first` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 139.
E.g. `(150, 150, 3)` would be one valid value.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
Returns:
A Keras model instance.
Raises:
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
"""
if weights not in {'imagenet', None}:
raise ValueError('The `weights` argument should be either '
'`None` (random initialization) or `imagenet` '
'(pre-training on ImageNet).')
# if weights == 'imagenet' and include_top and classes != 1001:
# raise ValueError('If using `weights` as imagenet with `include_top`'
# ' as true, `classes` should be 1001')
# Determine proper input shape
input_shape = _obtain_input_shape(
input_shape,
#default_size=299,
default_size=224,
min_size=139,
data_format=K.image_data_format(),
include_top=include_top)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
img_input = Input(tensor=input_tensor, shape=input_shape)
if K.image_data_format() == 'channels_first':
channel_axis = 1
else:
channel_axis = 3
# 'Sequential bit at start'
x = img_input
x = conv2d_bn(x, 64, 7, 7, strides=(2, 2), padding='same', name='Conv2d_1a_7x7')
x = MaxPooling2D((3, 3), strides=(2, 2), padding='same', name='MaxPool_2a_3x3')(x)
x = conv2d_bn(x, 64, 1, 1, strides=(1, 1), padding='same', name='Conv2d_2b_1x1')
x = conv2d_bn(x, 192, 3, 3, strides=(1, 1), padding='same', name='Conv2d_2c_3x3')
x = MaxPooling2D((3, 3), strides=(2, 2), padding='same', name='MaxPool_3a_3x3')(x)
# Now the '3' level inception units
x = concatenated_block(x, (( 64,), ( 96,128), (16, 32), ( 32,)), channel_axis, 'Mixed_3b')
x = concatenated_block(x, ((128,), (128,192), (32, 96), ( 64,)), channel_axis, 'Mixed_3c')
x = MaxPooling2D((3, 3), strides=(2, 2), padding='same', name='MaxPool_4a_3x3')(x)
# Now the '4' level inception units
x = concatenated_block(x, ((192,), ( 96,208), (16, 48), ( 64,)), channel_axis, 'Mixed_4b')
x = concatenated_block(x, ((160,), (112,224), (24, 64), ( 64,)), channel_axis, 'Mixed_4c')
x = concatenated_block(x, ((128,), (128,256), (24, 64), ( 64,)), channel_axis, 'Mixed_4d')
x = concatenated_block(x, ((112,), (144,288), (32, 64), ( 64,)), channel_axis, 'Mixed_4e')
x = concatenated_block(x, ((256,), (160,320), (32,128), (128,)), channel_axis, 'Mixed_4f')
x = MaxPooling2D((2, 2), strides=(2, 2), padding='same', name='MaxPool_5a_2x2')(x)
# Now the '5' level inception units
x = concatenated_block(x, ((256,), (160,320), (32,128), (128,)), channel_axis, 'Mixed_5b')
x = concatenated_block(x, ((384,), (192,384), (48,128), (128,)), channel_axis, 'Mixed_5c')
if include_top:
# Classification block
# 'AvgPool_0a_7x7'
x = AveragePooling2D((7, 7), strides=(1, 1), padding='valid')(x)
# 'Dropout_0b'
x = Dropout(0.5)(x) # slim has keep_prob (@0.8), keras uses drop_fraction
#logits = conv2d_bn(x, classes+1, 1, 1, strides=(1, 1), padding='valid', name='Logits',
# normalizer=False, activation=None, )
# Write out the logits explictly, since it is pretty different
x = Conv2D(classes, (1, 1), strides=(1,1), padding='valid', use_bias=True, name='Logits')(x)
#x = Flatten(name='Logits_flat')(x)
#print(x.get_shape().as_list())
#x = x[:, 1:] # ??Shift up so that first class ('blank background') vanishes
# Would be more efficient to strip off position[0] from the weights+bias terms directly in 'Logits'
#x = Conv2DTranspose(2, (64, 64), strides=(32, 32), filters=x.get_shape().as_list(), padding='same')
#x = Conv2DTranspose(2, (64, 64), strides=(32, 32), activation='softmax', padding='same')
x = BilinearUpSampling2D(target_size=tuple(image_size))(x)
#x = Conv2D(filters=2,
# kernel_size=(1, 1))(x)
#x = Conv2DTranspose(filters=2,
# kernel_size=(64, 64),
# strides=(32, 32),
# padding='same',
# activation='softmax',
#kernel_initializer=Constant(bilinear_upsample_weights(32, 2))
# )(x)
print(x.shape)
x = Flatten(name='Logits_flat')(x)
x = Activation('softmax', name='Predictions')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D(name='global_pooling')(x)
elif pooling == 'max':
x = GlobalMaxPooling2D(name='global_pooling')(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Finally : Create model
model = Model(inputs, x, name='inception_v1')
# LOAD model weights
if weights == 'imagenet':
if K.image_data_format() == 'channels_first':
if K.backend() == 'tensorflow':
warnings.warn('You are using the TensorFlow backend, yet you '
'are using the Theano '
'image data format convention '
'(`image_data_format="channels_first"`). '
'For best performance, set '
'`image_data_format="channels_last"` in '
'your Keras config '
'at ~/.keras/keras.json.')
if include_top:
weights_path ='/home/wli/Downloads/googlenet0917-02-0.93.hdf5'
#get_file(
# 'inception_v1_weights_tf_dim_ordering_tf_kernels.h5',
# WEIGHTS_PATH,
# cache_subdir='models',
# md5_hash='723bf2f662a5c07db50d28c8d35b626d')
else:
weights_path ='/home/wli/Downloads/googlenet0917-02-0.93.hdf5'
#weights_path = get_file(
# 'inception_v1_weights_tf_dim_ordering_tf_kernels_notop.h5',
# WEIGHTS_PATH_NO_TOP,
# cache_subdir='models',
# md5_hash='6fa8ecdc5f6c402a59909437f0f5c975')
model.load_weights('/home/wli/Downloads/googlenet0917-02-0.93.hdf5')
if K.backend() == 'theano':
convert_all_kernels_in_model(model)
return model
for (i,layer) in enumerate(ResNet50.layers):
print(str(i) + " "+ layer.__class__.__name__, layer.trainable)
def addTopModel(bottom_model, num_class):
my_model = bottom_model.output
my_model = GlobalAveragePooling2D()(my_model)
my_model = Dense(1024,activation='relu')(my_model)
my_model = Dense(1024,activation='relu')(my_model)
my_model = Dense(1024,activation='relu')(my_model)
my_model = Dense(512,activation='relu')(my_model)
my_model = Dense(num_classes,activation='softmax')(my_model)
return my_model
import cv2.cv2 as cv2, numpy as np
from keras.applications import DenseNet121
from keras.layers import Dense, GlobalAveragePooling2D
from keras import Model
IMG_PATH = ['./chest_xray_images/normal/15268.jpg', '0']
IMG_SHAPE = (320, 320, 3)
test_img = load_img(path=IMG_PATH[0], color_mode='grayscale')
test_img = img_to_array(img=test_img, data_format='channels_last')
test_img = cv2.resize(test_img, dsize=IMG_SHAPE[:2], interpolation=cv2.INTER_NEAREST)
test_img = np.expand_dims(test_img, axis=-1)
test_img = test_img.astype(np.uint8)
test_img = test_img / 255.
test_img = np.concatenate((test_img, test_img, test_img), axis=-1)
print('external image(s) shape:', test_img.shape)
backbone = DenseNet121(include_top=False, weights=None, input_shape=(320, 320, 3))
backbone_out = backbone.output
gap = GlobalAveragePooling2D(name='pooling_layer')(backbone_out)
output = Dense(units=14, activation='softmax', name='output_layer')(gap)
chexnet_model = Model(inputs=backbone.input, outputs=output)
chexnet_model.summary()
chexnet_model.load_weights('C:/Users/Arman/Desktop/Covid19-Detection/checkpoints/CheXNet/CheXNet_v0.3.0.h5')
chexnet_model.compile(optimizer='adam', loss='binary_crossentropy')
chexnet_model.save(filepath='./checkpoints/CheXNet/CheXNet_model.hdf5')
print('sample prediction: \n', chexnet_model.predict(np.expand_dims(test_img, axis=0)))
def se_block(in_block, ch, ratio=8):
z = GlobalAveragePooling2D()(in_block)
x = Dense(ch//ratio, activation='relu')(z)
x = Dense(ch, activation='sigmoid')(x)
return Multiply()([in_block, x])
def TS_ConvMobileNet(input_shape=None,
alpha=1.0,
depth_multiplier=1,
dropout=1e-3,
pooling=None,
classes=1000):
rows = None
# filter_num = [32,64,128,256,512,1024]
filter_num = [16, 32, 64, 128, 256, 512]
if input_shape:
if len(input_shape
) == 4 and input_shape[0] == 128 and input_shape[3] in [1, 3]:
img_input = Input(shape=input_shape)
else:
print('Please check the entered input shape.\n',
'The first input (timestep) must be 128\n',
'The last input must in 1 or 3')
raise ValueError
else:
img_input = Input(shape=(128, 64, 64, 1))
x = T_conv_block(img_input, filter_num[0], alpha, strides=(2, 2))
x = T_depthwise_conv_block(x,
filter_num[1],
alpha,
depth_multiplier,
block_id=1)
x = T_depthwise_conv_block(x,
filter_num[2],
alpha,
depth_multiplier,
strides=(2, 2),
block_id=2)
x = T_depthwise_conv_block(x,
filter_num[2],
alpha,
depth_multiplier,
block_id=3)
x = T_depthwise_conv_block(x,
filter_num[3],
alpha,
depth_multiplier,
strides=(2, 2),
block_id=4)
x = T_depthwise_conv_block(x,
filter_num[3],
alpha,
depth_multiplier,
block_id=5)
x = T_depthwise_conv_block(x,
filter_num[4],
alpha,
depth_multiplier,
strides=(2, 2),
block_id=6)
x = T_depthwise_conv_block(x,
filter_num[4],
alpha,
depth_multiplier,
block_id=7)
x = T_depthwise_conv_block(x,
filter_num[4],
alpha,
depth_multiplier,
block_id=8)
x = T_depthwise_conv_block(x,
filter_num[4],
alpha,
depth_multiplier,
block_id=9)
x = T_depthwise_conv_block(x,
filter_num[4],
alpha,
depth_multiplier,
block_id=10)
x = T_depthwise_conv_block(x,
filter_num[4],
alpha,
depth_multiplier,
block_id=11)
x = T_depthwise_conv_block(x,
filter_num[5],
alpha,
depth_multiplier,
strides=(2, 2),
block_id=12)
x = T_depthwise_conv_block(x,
filter_num[5],
alpha,
depth_multiplier,
block_id=13)
if backend.image_data_format() == 'channels_first':
shape = (1, int(filter_num[5] * alpha))
channelAxis = 1
else:
shape = (int(filter_num[5] * alpha), 1)
channelAxis = -1
if pooling == 'max':
x = TimeDistributed(GlobalMaxPooling2D(), name="MAX_pool")(x)
else:
x = TimeDistributed(GlobalAveragePooling2D(), name="AVG_pool")(x)
x = TimeDistributed(Reshape(shape, name='reshape_1'), name="reshape1")(x)
# TODO: Conv
# print(channelAxis)
TimeFilter_num = [256, 256, 256]
filter_step = [2, 16, 32]
conv1 = Conv2D(TimeFilter_num[0],
kernel_size=(filter_step[0], int(filter_num[5] * alpha)),
padding='valid',
name='time_conv_%d_step' % filter_step[0])(x)
conv1 = BatchNormalization(axis=channelAxis,
name='time_bn_%d_step' % filter_step[0])(conv1)
conv1 = Activation(backend.relu,
name='time_relu_%d_step' % filter_step[0])(conv1)
conv1 = MaxPool2D(pool_size=(128 - filter_step[0] + 1, 1),
strides=(1, 1),
padding='valid',
name='time_Max_%d_step' % filter_step[0])(conv1)
conv2 = Conv2D(TimeFilter_num[1],
kernel_size=(filter_step[1], int(filter_num[5] * alpha)),
padding='valid',
name='time_conv_%d_step' % filter_step[1])(x)
conv2 = BatchNormalization(axis=channelAxis,
name='time_bn_%d_step' % filter_step[1])(conv2)
conv2 = Activation(backend.relu,
name='time_relu_%d_step' % filter_step[1])(conv2)
conv2 = MaxPool2D(pool_size=(128 - filter_step[1] + 1, 1),
strides=(1, 1),
padding='valid',
name='time_Max_%d_step' % filter_step[1])(conv2)
conv3 = Conv2D(TimeFilter_num[2],
kernel_size=(filter_step[2], int(filter_num[5] * alpha)),
padding='valid',
name='time_conv_%d_step' % filter_step[2])(x)
conv3 = BatchNormalization(axis=channelAxis,
name='time_bn_%d_step' % filter_step[2])(conv3)
conv3 = Activation(backend.relu,
name='time_relu_%d_step' % filter_step[2])(conv3)
conv3 = MaxPool2D(pool_size=(128 - filter_step[2] + 1, 1),
strides=(1, 1),
padding='valid',
name='time_Max_%d_step' % filter_step[2])(conv3)
concatenated_tensor = Concatenate(axis=1)([conv1, conv2, conv3])
x = Flatten(name='flatten_concat')(concatenated_tensor)
x = Dropout(dropout)(x)
x = Dense(units=classes, activation='softmax')(x)
# get inputs shape
inputs = img_input
print(inputs)
# Create model.
model = models.Model(inputs, x, name='mobilenet_%0.2f_%s' % (alpha, rows))
return model
def ResNet152(include_top=True,
weights=None,
input_tensor=None,
input_shape=None,
large_input=False,
pooling=None,
classes=1000):
'''Instantiate the ResNet152 architecture.
# Arguments
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization)
or "imagenet" (pre-training on ImageNet).
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or `(3, 224, 224)` (with `channels_first` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 197.
E.g. `(200, 200, 3)` would be one valid value.
large_input: if True, then the input shape expected will be
`(448, 448, 3)` (with `channels_last` data format) or
`(3, 448, 448)` (with `channels_first` data format).
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
'''
if weights not in {'imagenet', None}:
raise ValueError('The `weights` argument should be either '
'`None` (random initialization) or `imagenet` '
'(pre-training on ImageNet).')
if weights == 'imagenet' and include_top and classes != 1000:
raise ValueError('If using `weights` as imagenet with `include_top`'
' as true, `classes` should be 1000')
eps = 1.1e-5
'''
if large_input:
img_size = 448
else:
img_size = 224
'''
img_size = 448
# Determine proper input shape
input_shape = _obtain_input_shape(input_shape,
default_size=img_size,
min_size=197,
data_format=K.image_data_format(),
include_top=include_top)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
# handle dimension ordering for different backends
if K.image_dim_ordering() == 'tf':
bn_axis = 3
else:
bn_axis = 1
x = ZeroPadding2D((3, 3), name='conv1_zeropadding')(img_input)
x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1', use_bias=False)(x)
x = BatchNormalization(epsilon=eps, axis=bn_axis, name='bn_conv1')(x)
x = Scale(axis=bn_axis, name='scale_conv1')(x)
x = Activation('relu', name='conv1_relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2), name='pool1')(x)
x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
x = conv_block(x, 3, [128, 128, 512], stage=3, block='a')
for i in range(1, 8):
x = identity_block(x, 3, [128, 128, 512], stage=3, block='b' + str(i))
x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a')
for i in range(1, 36):
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b' + str(i))
x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c')
if large_input:
x = AveragePooling2D((14, 14), name='avg_pool')(x)
else:
x = AveragePooling2D((7, 7), name='avg_pool')(x)
# include classification layer by default, not included for feature extraction
if include_top:
x = Flatten()(x)
x = Dense(classes, activation='softmax', name='fc1000')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model.
model = Model(inputs, x, name='resnet152')
# load weights
if weights == 'imagenet':
if include_top:
weights_path = get_file(
'resnet152_weights_tf.h5',
WEIGHTS_PATH,
cache_subdir='models',
md5_hash='cdb18a2158b88e392c0905d47dcef965')
else:
weights_path = get_file(
'resnet152_weights_tf_notop.h5',
WEIGHTS_PATH_NO_TOP,
cache_subdir='models',
md5_hash='4a90dcdafacbd17d772af1fb44fc2660')
model.load_weights(weights_path, by_name=True)
if K.backend() == 'theano':
layer_utils.convert_all_kernels_in_model(model)
if include_top:
maxpool = model.get_layer(name='avg_pool')
shape = maxpool.output_shape[1:]
dense = model.get_layer(name='fc1000')
layer_utils.convert_dense_weights_data_format(
dense, shape, 'channels_first')
if K.image_data_format() == 'channels_first' and K.backend(
) == 'tensorflow':
warnings.warn('You are using the TensorFlow backend, yet you '
'are using the Theano '
'image data format convention '
'(`image_data_format="channels_first"`). '
'For best performance, set '
'`image_data_format="channels_last"` in '
'your Keras config '
'at ~/.keras/keras.json.')
return model
def createModel_cqt_classification_fma_medium(input_dim, input_dim2,
selected_optimizer,
selected_loss):
fixed_init = he_normal(seed=0)
print('classifier model')
input = Input(shape=(1, input_dim, input_dim2)) #1 x 80 x 1280
out1 = Convolution2D(32, (3, 3),
activation='elu',
padding='same',
data_format='channels_first',
kernel_initializer=fixed_init)(input) #32 x 80 x 1280
out1 = Dropout(rate=0.1)(out1)
out1 = BatchNormalization(axis=1)(out1)
out1 = MaxPooling2D((2, 2), padding='same',
data_format='channels_first')(out1) #32 x 40 x 640
out2 = Convolution2D(32, (3, 3),
activation='elu',
padding='same',
data_format='channels_first',
kernel_initializer=fixed_init)(out1) #32 x 40 x 640
out2 = Dropout(rate=0.1)(out2)
out2 = BatchNormalization(axis=1)(out2)
out2 = MaxPooling2D((2, 2), padding='same',
data_format='channels_first')(out2) #32 x 20 x 320
out3 = Convolution2D(32, (3, 3),
activation='elu',
padding='same',
data_format='channels_first',
kernel_initializer=fixed_init)(out2) #32 x 20 x 320
out3 = Dropout(rate=0.1)(out3)
out3 = BatchNormalization(axis=1)(out3)
out3 = MaxPooling2D((2, 2), padding='same',
data_format='channels_first')(out3) #32 x 10 x 160
out4 = Convolution2D(32, (3, 3),
activation='elu',
padding='same',
data_format='channels_first',
kernel_initializer=fixed_init)(out3) #32 x 10 x 160
out4 = Dropout(rate=0.1)(out4)
out4 = BatchNormalization(axis=1)(out4)
out4 = MaxPooling2D((2, 2), padding='same',
data_format='channels_first')(out4) #32 x 5 x 80
out5 = Convolution2D(32, (3, 3),
activation='elu',
padding='same',
data_format='channels_first',
kernel_initializer=fixed_init)(out4) #32 x 5 x 80
out5 = Dropout(rate=0.1)(out5)
out5 = BatchNormalization(axis=1)(out5)
out5 = MaxPooling2D((5, 5), padding='same',
data_format='channels_first')(out5) #32 x 1 x 16
out6 = GlobalAveragePooling2D(data_format='channels_first')(
out5) #same as previous
output = Dense(16, activation='softmax')(out6)
#==== create model
classifier = Model(input, output)
layer1_extractor = Model(input, out1)
layer2_extractor = Model(input, out2)
layer3_extractor = Model(input, out3)
layer4_extractor = Model(input, out4)
layer5_extractor = Model(input, out5)
#layer6_extractor = Model(input, out4i)
#==== compile model
classifier.compile(optimizer=selected_optimizer,
loss=selected_loss,
metrics=['acc'])
return classifier, layer1_extractor, layer2_extractor, layer3_extractor, layer4_extractor, layer5_extractor
def TempInceptionV3(include_top=True,
weights='imagenet',
input_tensor=None,
input_shape=None,
pooling=None,
classes=1000,
depth=20,
pre='custom_'):
# Determine proper input shape
if input_shape is None:
input_shape = (None,None,depth)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
if K.image_data_format() == 'channels_first':
channel_axis = 1
else:
channel_axis = 3
x = Lambda(regulization, output_shape=output_of_lambda)(img_input)
x = conv2d_bn(x, 32, 3, 3, strides=(2, 2), padding='valid')
x = conv2d_bn(x, 32, 3, 3, padding='valid')
x = conv2d_bn(x, 64, 3, 3)
x = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = conv2d_bn(x, 80, 1, 1, padding='valid')
x = conv2d_bn(x, 192, 3, 3, padding='valid')
x = MaxPooling2D((3, 3), strides=(2, 2))(x)
# mixed 0, 1, 2: 35 x 35 x 256
branch1x1 = conv2d_bn(x, 64, 1, 1)
branch5x5 = conv2d_bn(x, 48, 1, 1)
branch5x5 = conv2d_bn(branch5x5, 64, 5, 5)
branch3x3dbl = conv2d_bn(x, 64, 1, 1)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 32, 1, 1)
x = layers.concatenate(
[branch1x1, branch5x5, branch3x3dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed0')
# mixed 1: 35 x 35 x 256
branch1x1 = conv2d_bn(x, 64, 1, 1)
branch5x5 = conv2d_bn(x, 48, 1, 1)
branch5x5 = conv2d_bn(branch5x5, 64, 5, 5)
branch3x3dbl = conv2d_bn(x, 64, 1, 1)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 64, 1, 1)
x = layers.concatenate(
[branch1x1, branch5x5, branch3x3dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed1')
# mixed 2: 35 x 35 x 256
branch1x1 = conv2d_bn(x, 64, 1, 1)
branch5x5 = conv2d_bn(x, 48, 1, 1)
branch5x5 = conv2d_bn(branch5x5, 64, 5, 5)
branch3x3dbl = conv2d_bn(x, 64, 1, 1)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 64, 1, 1)
x = layers.concatenate(
[branch1x1, branch5x5, branch3x3dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed2')
# mixed 3: 17 x 17 x 768
branch3x3 = conv2d_bn(x, 384, 3, 3, strides=(2, 2), padding='valid')
branch3x3dbl = conv2d_bn(x, 64, 1, 1)
branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
branch3x3dbl = conv2d_bn(
branch3x3dbl, 96, 3, 3, strides=(2, 2), padding='valid')
branch_pool = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = layers.concatenate(
[branch3x3, branch3x3dbl, branch_pool], axis=channel_axis, name=pre + 'mixed3')
#loss1
# mixed 4: 17 x 17 x 768
branch1x1 = conv2d_bn(x, 192, 1, 1)
branch7x7 = conv2d_bn(x, 128, 1, 1)
branch7x7 = conv2d_bn(branch7x7, 128, 1, 7)
branch7x7 = conv2d_bn(branch7x7, 192, 7, 1)
branch7x7dbl = conv2d_bn(x, 128, 1, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 128, 7, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 128, 1, 7)
branch7x7dbl = conv2d_bn(branch7x7dbl, 128, 7, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
x = layers.concatenate(
[branch1x1, branch7x7, branch7x7dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed4')
# mixed 5, 6: 17 x 17 x 768
for i in range(2):
branch1x1 = conv2d_bn(x, 192, 1, 1)
branch7x7 = conv2d_bn(x, 160, 1, 1)
branch7x7 = conv2d_bn(branch7x7, 160, 1, 7)
branch7x7 = conv2d_bn(branch7x7, 192, 7, 1)
branch7x7dbl = conv2d_bn(x, 160, 1, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 160, 7, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 160, 1, 7)
branch7x7dbl = conv2d_bn(branch7x7dbl, 160, 7, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
branch_pool = AveragePooling2D(
(3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
x = layers.concatenate(
[branch1x1, branch7x7, branch7x7dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed' + str(5 + i))
# mixed 7: 17 x 17 x 768
branch1x1 = conv2d_bn(x, 192, 1, 1)
branch7x7 = conv2d_bn(x, 192, 1, 1)
branch7x7 = conv2d_bn(branch7x7, 192, 1, 7)
branch7x7 = conv2d_bn(branch7x7, 192, 7, 1)
branch7x7dbl = conv2d_bn(x, 192, 1, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 7, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 7, 1)
branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
branch_pool = AveragePooling2D((3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
x = layers.concatenate(
[branch1x1, branch7x7, branch7x7dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed7')
# mixed 8: 8 x 8 x 1280
branch3x3 = conv2d_bn(x, 192, 1, 1)
branch3x3 = conv2d_bn(branch3x3, 320, 3, 3,
strides=(2, 2), padding='valid')
branch7x7x3 = conv2d_bn(x, 192, 1, 1)
branch7x7x3 = conv2d_bn(branch7x7x3, 192, 1, 7)
branch7x7x3 = conv2d_bn(branch7x7x3, 192, 7, 1)
branch7x7x3 = conv2d_bn(
branch7x7x3, 192, 3, 3, strides=(2, 2), padding='valid')
branch_pool = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = layers.concatenate(
[branch3x3, branch7x7x3, branch_pool], axis=channel_axis, name=pre + 'mixed8')
#loss2
# mixed 9: 8 x 8 x 2048
for i in range(2):
branch1x1 = conv2d_bn(x, 320, 1, 1)
branch3x3 = conv2d_bn(x, 384, 1, 1)
branch3x3_1 = conv2d_bn(branch3x3, 384, 1, 3)
branch3x3_2 = conv2d_bn(branch3x3, 384, 3, 1)
branch3x3 = layers.concatenate(
[branch3x3_1, branch3x3_2], axis=channel_axis, name=pre + 'mixed9_' + str(i))
branch3x3dbl = conv2d_bn(x, 448, 1, 1)
branch3x3dbl = conv2d_bn(branch3x3dbl, 384, 3, 3)
branch3x3dbl_1 = conv2d_bn(branch3x3dbl, 384, 1, 3)
branch3x3dbl_2 = conv2d_bn(branch3x3dbl, 384, 3, 1)
branch3x3dbl = layers.concatenate(
[branch3x3dbl_1, branch3x3dbl_2], axis=channel_axis)
branch_pool = AveragePooling2D(
(3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
x = layers.concatenate(
[branch1x1, branch3x3, branch3x3dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed' + str(9 + i))
if include_top:
# Classification block
x = GlobalAveragePooling2D(name=pre + 'avg_pool')(x)
x = Dense(classes, activation='softmax', name=pre + 'predictions')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
inputs = img_input
# Create model.
model = Model(inputs, x, name=pre + 'inception_v3')
# load weights
if weights == 'imagenet':
model.load_weights('data/inceptionv3_temporal.h5')
print ('Loaded weights')
return model
filters_3x3=320,
filters_5x5_reduce=32,
filters_5x5=128,
filters_pool_proj=128,
name='inception_5a')
x = inception_module(x,
filters_1x1=384,
filters_3x3_reduce=192,
filters_3x3=384,
filters_5x5_reduce=48,
filters_5x5=128,
filters_pool_proj=128,
name='inception_5b')
x = GlobalAveragePooling2D(name='avg_pool_5_3x3/1')(x)
x = Dropout(0.4)(x)
x = Dense(38, activation='softmax', name='output')(x)
model = Model(input_layer, [x, x1, x2], name='inception_v1')
model.summary()
epochs = 25
initial_lrate = 0.005
def decay(epoch, steps=100):
initial_lrate = 0.005
drop = 0.96
def InceptionResNetV2(include_top=True,
weights='imagenet',
input_tensor=None,
input_shape=None,
pooling=None,
classes=1000):
"""Instantiates the Inception-ResNet v2 architecture.
Optionally loads weights pre-trained on ImageNet.
Note that when using TensorFlow, for best performance you should
set `"image_data_format": "channels_last"` in your Keras config
at `~/.keras/keras.json`.
The model and the weights are compatible with TensorFlow, Theano and
CNTK backends. The data format convention used by the model is
the one specified in your Keras config file.
Note that the default input image size for this model is 299x299, instead
of 224x224 as in the VGG16 and ResNet models. Also, the input preprocessing
function is different (i.e., do not use `imagenet_utils.preprocess_input()`
with this model. Use `preprocess_input()` defined in this module instead).
# Arguments
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization),
'imagenet' (pre-training on ImageNet),
or the path to the weights file to be loaded.
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is `False` (otherwise the input shape
has to be `(299, 299, 3)` (with `'channels_last'` data format)
or `(3, 299, 299)` (with `'channels_first'` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 139.
E.g. `(150, 150, 3)` would be one valid value.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the last convolutional layer.
- `'avg'` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `'max'` means that global max pooling will be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is `True`, and
if no `weights` argument is specified.
# Returns
A Keras `Model` instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
"""
if not (weights in {'imagenet', None} or os.path.exists(weights)):
raise ValueError('The `weights` argument should be either '
'`None` (random initialization), `imagenet` '
'(pre-training on ImageNet), '
'or the path to the weights file to be loaded.')
if weights == 'imagenet' and include_top and classes != 1000:
raise ValueError('If using `weights` as imagenet with `include_top`'
' as true, `classes` should be 1000')
# Determine proper input shape
input_shape = _obtain_input_shape(input_shape,
default_size=299,
min_size=139,
data_format=K.image_data_format(),
require_flatten=False,
weights=weights)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
# Stem block: 35 x 35 x 192
x = conv2d_bn(img_input, 32, 3, strides=2, padding='same')
x = conv2d_bn(x, 32, 3, padding='same')
x = conv2d_bn(x, 64, 3)
x = MaxPooling2D(3, strides=2, padding='same')(x)
x = conv2d_bn(x, 80, 1, padding='same')
x = conv2d_bn(x, 192, 3, padding='same')
x = MaxPooling2D(3, strides=2, padding='same')(x)
# Mixed 5b (Inception-A block): 35 x 35 x 320
branch_0 = conv2d_bn(x, 96, 1)
branch_1 = conv2d_bn(x, 48, 1)
branch_1 = conv2d_bn(branch_1, 64, 5)
branch_2 = conv2d_bn(x, 64, 1)
branch_2 = conv2d_bn(branch_2, 96, 3)
branch_2 = conv2d_bn(branch_2, 96, 3)
branch_pool = AveragePooling2D(3, strides=1, padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 64, 1)
branches = [branch_0, branch_1, branch_2, branch_pool]
channel_axis = 1 if K.image_data_format() == 'channels_first' else 3
x = Concatenate(axis=channel_axis, name='mixed_5b')(branches)
# 10x block35 (Inception-ResNet-A block): 35 x 35 x 320
for block_idx in range(1, 11):
x = inception_resnet_block(x,
scale=0.17,
block_type='block35',
block_idx=block_idx)
# Mixed 6a (Reduction-A block): 17 x 17 x 1088
branch_0 = conv2d_bn(x, 384, 3, strides=2, padding='same')
branch_1 = conv2d_bn(x, 256, 1)
branch_1 = conv2d_bn(branch_1, 256, 3)
branch_1 = conv2d_bn(branch_1, 384, 3, strides=2, padding='same')
branch_pool = MaxPooling2D(3, strides=2, padding='same')(x)
branches = [branch_0, branch_1, branch_pool]
x = Concatenate(axis=channel_axis, name='mixed_6a')(branches)
# 20x block17 (Inception-ResNet-B block): 17 x 17 x 1088
for block_idx in range(1, 21):
x = inception_resnet_block(x,
scale=0.1,
block_type='block17',
block_idx=block_idx)
# Mixed 7a (Reduction-B block): 8 x 8 x 2080
branch_0 = conv2d_bn(x, 256, 1)
branch_0 = conv2d_bn(branch_0, 384, 3, strides=2, padding='same')
branch_1 = conv2d_bn(x, 256, 1)
branch_1 = conv2d_bn(branch_1, 288, 3, strides=2, padding='same')
branch_2 = conv2d_bn(x, 256, 1)
branch_2 = conv2d_bn(branch_2, 288, 3)
branch_2 = conv2d_bn(branch_2, 320, 3, strides=2, padding='same')
branch_pool = MaxPooling2D(3, strides=2, padding='same')(x)
branches = [branch_0, branch_1, branch_2, branch_pool]
x = Concatenate(axis=channel_axis, name='mixed_7a')(branches)
# 10x block8 (Inception-ResNet-C block): 8 x 8 x 2080
for block_idx in range(1, 10):
x = inception_resnet_block(x,
scale=0.2,
block_type='block8',
block_idx=block_idx)
x = inception_resnet_block(x,
scale=1.,
activation=None,
block_type='block8',
block_idx=10)
# Final convolution block: 8 x 8 x 1536
x = conv2d_bn(x, 1536, 1, name='conv_7b')
if include_top:
# Classification block
x = GlobalAveragePooling2D(name='avg_pool')(x)
x = Dense(classes, activation='softmax', name='predictions')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model
model = Model(inputs, x, name='inception_resnet_v2')
# Load weights
if weights == 'imagenet':
if K.image_data_format() == 'channels_first':
if K.backend() == 'tensorflow':
warnings.warn('You are using the TensorFlow backend, yet you '
'are using the Theano '
'image data format convention '
'(`image_data_format="channels_first"`). '
'For best performance, set '
'`image_data_format="channels_last"` in '
'your Keras config '
'at ~/.keras/keras.json.')
if include_top:
fname = 'inception_resnet_v2_weights_tf_dim_ordering_tf_kernels.h5'
weights_path = get_file(
fname,
BASE_WEIGHT_URL + fname,
cache_subdir='/wdata/backbones_weights',
file_hash='e693bd0210a403b3192acc6073ad2e96')
else:
fname = 'inception_resnet_v2_weights_tf_dim_ordering_tf_kernels_notop.h5'
weights_path = get_file(
fname,
BASE_WEIGHT_URL + fname,
cache_subdir='/wdata/backbones_weights',
file_hash='d19885ff4a710c122648d3b5c3b684e4')
model.load_weights(weights_path, skip_mismatch=True)
elif weights is not None:
model.load_weights(weights)
return model
def Inception_v3c(input_tensor=None, input_shape=None, pre='custom_', weights='imagenet', depth=3, pooling='avg'):
if input_shape is None:
input_shape = (None,None,depth)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
if K.image_data_format() == 'channels_first':
channel_axis = 1
else:
channel_axis = 3
x = Input(tensor=img_input)
# mixed 8: 8 x 8 x 1280
branch3x3 = conv2d_bn(x, 192, 1, 1)
branch3x3 = conv2d_bn(branch3x3, 320, 3, 3,
strides=(2, 2), padding='valid')
branch7x7x3 = conv2d_bn(x, 192, 1, 1)
branch7x7x3 = conv2d_bn(branch7x7x3, 192, 1, 7)
branch7x7x3 = conv2d_bn(branch7x7x3, 192, 7, 1)
branch7x7x3 = conv2d_bn(
branch7x7x3, 192, 3, 3, strides=(2, 2), padding='valid')
branch_pool = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = layers.concatenate(
[branch3x3, branch7x7x3, branch_pool], axis=channel_axis, name=pre + 'mixed8')
# mixed 9: 8 x 8 x 2048
for i in range(2):
branch1x1 = conv2d_bn(x, 320, 1, 1)
branch3x3 = conv2d_bn(x, 384, 1, 1)
branch3x3_1 = conv2d_bn(branch3x3, 384, 1, 3)
branch3x3_2 = conv2d_bn(branch3x3, 384, 3, 1)
branch3x3 = layers.concatenate(
[branch3x3_1, branch3x3_2], axis=channel_axis, name=pre + 'mixed9_' + str(i))
branch3x3dbl = conv2d_bn(x, 448, 1, 1)
branch3x3dbl = conv2d_bn(branch3x3dbl, 384, 3, 3)
branch3x3dbl_1 = conv2d_bn(branch3x3dbl, 384, 1, 3)
branch3x3dbl_2 = conv2d_bn(branch3x3dbl, 384, 3, 1)
branch3x3dbl = layers.concatenate(
[branch3x3dbl_1, branch3x3dbl_2], axis=channel_axis)
branch_pool = AveragePooling2D(
(3, 3), strides=(1, 1), padding='same')(x)
branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
x = layers.concatenate(
[branch1x1, branch3x3, branch3x3dbl, branch_pool],
axis=channel_axis,
name=pre + 'mixed' + str(9 + i))
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
inputs = img_input
# Create model.
model = Model(inputs, x, name=pre + 'inception_v3c')
if weights == 'imagenet':
model.load_weights('data/InceptionV3c.h5')
print ('Loaded weights v3c')
return model
def ResNet50(include_top=True,
weights='imagenet',
input_tensor=None,
input_shape=None,
pooling=None,
classes=1000):
"""Instantiates the ResNet50 architecture.
Optionally loads weights pre-trained
on ImageNet. Note that when using TensorFlow,
for best performance you should set
`image_data_format="channels_last"` in your Keras config
at ~/.keras/keras.json.
The model and the weights are compatible with both
TensorFlow and Theano. The data format
convention used by the model is the one
specified in your Keras config file.
# Arguments
include_top: whether to include the 3 fully-connected
layers at the top of the network.
weights: one of `None` (random initialization)
or "imagenet" (pre-training on ImageNet).
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or `(3, 224, 244)` (with `channels_first` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 197.
E.g. `(200, 200, 3)` would be one valid value.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
"""
if weights not in {'imagenet', None}:
raise ValueError('The `weights` argument should be either '
'`None` (random initialization) or `imagenet` '
'(pre-training on ImageNet).')
if weights == 'imagenet' and include_top and classes != 1000:
raise ValueError('If using `weights` as imagenet with `include_top`'
' as true, `classes` should be 1000')
# Determine proper input shape
input_shape = _obtain_input_shape(input_shape,
default_size=224,
min_size=197,
data_format=K.image_data_format(),
require_flatten=include_top,
weights=weights)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
if K.image_data_format() == 'channels_last':
bn_axis = 3
else:
bn_axis = 1
x = ZeroPadding2D((3, 3))(img_input)
x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1')(x)
x = BatchNormalization(axis=bn_axis, name='bn_conv1')(x)
x = Activation('relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
x = conv_block(x, 3, [128, 128, 512], stage=3, block='a')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='b')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='c')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='d')
x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='c')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='d')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='e')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='f')
x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c')
x = AveragePooling2D((7, 7), name='avg_pool')(x)
if include_top:
x = Flatten()(x)
x = Dense(classes, activation='softmax', name='fc1000')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model.
model = Model(inputs, x, name='resnet50')
# load weights
if weights == 'imagenet':
if include_top:
weights_path = get_file(
'resnet50_weights_tf_dim_ordering_tf_kernels.h5',
WEIGHTS_PATH,
cache_subdir='models',
md5_hash='a7b3fe01876f51b976af0dea6bc144eb')
else:
weights_path = get_file(
'resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5',
WEIGHTS_PATH_NO_TOP,
cache_subdir='models',
md5_hash='a268eb855778b3df3c7506639542a6af')
model.load_weights(weights_path)
if K.backend() == 'theano':
layer_utils.convert_all_kernels_in_model(model)
if K.image_data_format() == 'channels_first':
if include_top:
maxpool = model.get_layer(name='avg_pool')
shape = maxpool.output_shape[1:]
dense = model.get_layer(name='fc1000')
layer_utils.convert_dense_weights_data_format(
dense, shape, 'channels_first')
if K.backend() == 'tensorflow':
warnings.warn('You are using the TensorFlow backend, yet you '
'are using the Theano '
'image data format convention '
'(`image_data_format="channels_first"`). '
'For best performance, set '
'`image_data_format="channels_last"` in '
'your Keras config '
'at ~/.keras/keras.json.')
return model
def ResNet50(include_top=True,
weights='imagenet',
input_tensor=None,
input_shape=None,
output_shape=None,
dense_shape=None,
pooling=None,
classes=1000):
"""Instantiates the ResNet50 architecture.
Optionally loads weights pre-trained
on ImageNet. Note that when using TensorFlow,
for best performance you should set
`image_data_format="channels_last"` in your Keras config
at ~/.keras/keras.json.
The model and the weights are compatible with both
TensorFlow and Theano. The data format
convention used by the model is the one
specified in your Keras config file.
# Arguments
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization)
or "imagenet" (pre-training on ImageNet).
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or `(3, 224, 224)` (with `channels_first` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 197.
E.g. `(200, 200, 3)` would be one valid value.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
"""
# Determine proper input shape
input_shape = _obtain_input_shape(input_shape,
default_size=224,
min_size=197,
data_format=K.image_data_format(),
include_top=include_top)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
if K.image_data_format() == 'channels_last':
bn_axis = 3
else:
bn_axis = 1
x = ZeroPadding2D((3, 3))(img_input)
x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1')(x)
x = BatchNormalization(axis=bn_axis, name='bn_conv1')(x)
x = Activation('relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
x = conv_block(x, 3, [128, 128, 512], stage=3, block='a')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='b')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='c')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='d')
x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='c')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='d')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='e')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='f')
x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c')
x = AveragePooling2D((7, 7), name='avg_pool')(x)
if include_top:
x = Flatten()(x)
x = Dense(classes, activation='softmax', name='fc1000')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model
base_model = Model(inputs, x, name='resnet50')
for layer in base_model.layers:
layer.trainable = False
#print (model.summary())
#model = Model(inputs, x, name='resnet50')
return base_model
def DenseNet(blocks,
include_top=True,
weights='imagenet',
model_path=None,
input_tensor=None,
input_shape=(224, 224, 3),
pooling=None,
classes=1000):
"""Instantiates the DenseNet architecture.
Optionally loads weights pre-trained
on ImageNet. Note that when using TensorFlow,
for best performance you should set
`image_data_format='channels_last'` in your Keras config
at ~/.keras/keras.json.
The model and the weights are compatible with
TensorFlow, Theano, and CNTK. The data format
convention used by the model is the one
specified in your Keras config file.
# Arguments
blocks: numbers of building blocks for the four dense layers.
include_top: whether to include the fully-connected
layer at the top of the network.
weights: one of `None` (random initialization),
'imagenet' (pre-training on ImageNet),
or the path to the weights file to be loaded.
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or `(3, 224, 224)` (with `channels_first` data format).
It should have exactly 3 inputs channels.
pooling: optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
"""
if not (weights in {'imagenet', None} or os.path.exists(weights)):
raise ValueError('The `weights` argument should be either '
'`None` (random initialization), `imagenet` '
'(pre-training on ImageNet), '
'or the path to the weights file to be loaded.')
if weights == 'imagenet' and include_top and classes != 1000:
raise ValueError('If using `weights` as imagenet with `include_top`'
' as true, `classes` should be 1000')
# Determine proper input shape
# input_shape = _obtain_input_shape(input_shape,
# default_size=224,
# min_size=221,
# data_format=K.image_data_format(),
# require_flatten=include_top,
# weights=weights)
# input_shape = (224,224,3)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
bn_axis = 3 if K.image_data_format() == 'channels_last' else 1
x = ZeroPadding2D(padding=((3, 3), (3, 3)))(img_input)
x = Conv2D(64, 7, strides=2, use_bias=False, name='conv1/conv')(x)
x = BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name='conv1/bn')(x)
x = Activation('relu', name='conv1/relu')(x)
x = ZeroPadding2D(padding=((1, 1), (1, 1)))(x)
x = MaxPooling2D(3, strides=2, name='pool1')(x)
x = dense_block(x, blocks[0], name='conv2')
x = transition_block(x, 0.5, name='pool2')
x = dense_block(x, blocks[1], name='conv3')
x = transition_block(x, 0.5, name='pool3')
x = dense_block(x, blocks[2], name='conv4')
x = transition_block(x, 0.5, name='pool4')
x = dense_block(x, blocks[3], name='conv5')
x = BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name='bn')(x)
if include_top:
x = GlobalAveragePooling2D(name='avg_pool')(x)
x = Dense(classes, activation='softmax', name='fc1000')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D(name='avg_pool')(x)
elif pooling == 'max':
x = GlobalMaxPooling2D(name='max_pool')(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model.
if blocks == [6, 12, 24, 16]:
model = Model(inputs, x, name='densenet121')
elif blocks == [6, 12, 32, 32]:
model = Model(inputs, x, name='densenet169')
elif blocks == [6, 12, 48, 32]:
model = Model(inputs, x, name='densenet201')
else:
model = Model(inputs, x, name='densenet')
# Load weights.
if weights == 'imagenet':
if include_top:
if blocks == [6, 12, 24, 16]:
weights_path = get_file(
os.path.join(
model_path,
'densenet121_weights_tf_dim_ordering_tf_kernels.h5'),
DENSENET121_WEIGHT_PATH,
cache_subdir='./model_save',
file_hash='0962ca643bae20f9b6771cb844dca3b0')
elif blocks == [6, 12, 32, 32]:
weights_path = get_file(
os.path.join(
model_path,
'densenet169_weights_tf_dim_ordering_tf_kernels.h5'),
DENSENET169_WEIGHT_PATH,
cache_subdir='./model_save',
file_hash='bcf9965cf5064a5f9eb6d7dc69386f43')
elif blocks == [6, 12, 48, 32]:
weights_path = get_file(
os.path.join(
model_path,
'densenet201_weights_tf_dim_ordering_tf_kernels.h5'),
DENSENET201_WEIGHT_PATH,
cache_subdir='./model_save',
file_hash='7bb75edd58cb43163be7e0005fbe95ef')
else:
if blocks == [6, 12, 24, 16]:
weights_path = get_file(os.path.join(
model_path,
'densenet121_weights_tf_dim_ordering_tf_kernels_notop.h5'),
DENSENET121_WEIGHT_PATH_NO_TOP,
cache_subdir='./model_save',
file_hash=
'4912a53fbd2a69346e7f2c0b5ec8c6d3')
elif blocks == [6, 12, 32, 32]:
weights_path = get_file(os.path.join(
model_path,
'densenet169_weights_tf_dim_ordering_tf_kernels_notop.h5'),
DENSENET169_WEIGHT_PATH_NO_TOP,
cache_subdir='./model_save',
file_hash=
'50662582284e4cf834ce40ab4dfa58c6')
elif blocks == [6, 12, 48, 32]:
weights_path = get_file(os.path.join(
model_path,
'densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5'),
DENSENET201_WEIGHT_PATH_NO_TOP,
cache_subdir='./model_save',
file_hash=
'1c2de60ee40562448dbac34a0737e798')
model.load_weights(weights_path)
print("Loading weights from {}".format(weights_path))
elif weights is not None:
model.load_weights(weights)
return model
def VGG16(include_top=True,
weights=None,
input_tensor=None,
input_shape=None,
pooling=None,
nfcfilters=4096,
nconvfilters=[64, 128, 512, 512, 512],
classes=1,
nblocks=5):
seq_input = Input(shape=input_shape)
if nblocks >= 1:
# Block 1
x = Conv2D(nconvfilters[0], (4, 3),
activation='relu',
padding='same',
name='block1_conv1')(seq_input)
x = Conv2D(nconvfilters[0], (1, 3),
activation='relu',
padding='same',
name='block1_conv2')(x)
x = MaxPooling2D((1, 2), name='block1_pool')(x)
if nblocks >= 2:
# Block 2
x = Conv2D(nconvfilters[1], (1, 3),
activation='relu',
padding='same',
name='block2_conv1')(x)
x = Conv2D(nconvfilters[1], (1, 3),
activation='relu',
padding='same',
name='block2_conv2')(x)
x = MaxPooling2D((1, 2), name='block2_pool')(x)
if nblocks >= 3:
# Block 3
x = Conv2D(nconvfilters[2], (1, 3),
activation='relu',
padding='same',
name='block3_conv1')(x)
x = Conv2D(nconvfilters[2], (1, 3),
activation='relu',
padding='same',
name='block3_conv2')(x)
x = Conv2D(nconvfilters[2], (1, 3),
activation='relu',
padding='same',
name='block3_conv3')(x)
x = MaxPooling2D((1, 2), name='block3_pool')(x)
if nblocks >= 4:
# Block 4
x = Conv2D(nconvfilters[3], (1, 3),
activation='relu',
padding='same',
name='block4_conv1')(x)
x = Conv2D(nconvfilters[3], (1, 3),
activation='relu',
padding='same',
name='block4_conv2')(x)
x = Conv2D(nconvfilters[3], (1, 3),
activation='relu',
padding='same',
name='block4_conv3')(x)
x = MaxPooling2D((1, 2), name='block4_pool')(x)
if nblocks >= 5:
# Block 5
x = Conv2D(nconvfilters[4], (1, 3),
activation='relu',
padding='same',
name='block5_conv1')(x)
x = Conv2D(nconvfilters[4], (1, 3),
activation='relu',
padding='same',
name='block5_conv2')(x)
x = Conv2D(nconvfilters[4], (1, 3),
activation='relu',
padding='same',
name='block5_conv3')(x)
x = MaxPooling2D((1, 2), name='block5_pool')(x)
if include_top:
# Classification block
x = Flatten(name='flatten')(x)
x = Dense(nfcfilters, activation='relu', name='fc1')(x)
x = Dense(nfcfilters, activation='relu', name='fc2')(x)
if classes == 1:
x = Dense(classes, activation='sigmoid', name='predictions')(x)
else:
x = Dense(classes, activation='softmax', name='predictions')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# inputs
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = seq_input
# Create model
model = Model(inputs, x, name='vgg16')
if weights:
raise NotImplementedError
return model
def build(width,
height,
depth,
classes,
stages,
filters,
reg=0.0001,
bnEps=2e-5,
bnMom=0.9):
inputShape = (height, width, depth)
chanDim = -1
inputs = Input(shape=inputShape)
x = BatchNormalization(axis=chanDim,
epsilon=bnEps,
momentum=bnMom,
beta_initializer="zeros",
gamma_initializer="ones")(inputs)
x = Activation("relu")(x)
x = SeparableConv2D(64, (3, 3),
use_bias=False,
padding="same",
depthwise_regularizer=l2(reg),
depthwise_initializer='glorot_uniform')(x)
x = SeparableConv2D(128, (3, 3),
use_bias=False,
padding="same",
depthwise_regularizer=l2(reg),
depthwise_initializer='glorot_uniform')(x)
x = SeparableConv2D(256, (3, 3),
use_bias=False,
padding="same",
depthwise_regularizer=l2(reg),
depthwise_initializer='glorot_uniform')(x)
x = BatchNormalization(axis=chanDim,
epsilon=bnEps,
momentum=bnMom,
beta_initializer="zeros",
gamma_initializer="ones")(x)
x = Activation("relu")(x)
x = ZeroPadding2D((1, 1))(x)
x = MaxPooling2D((3, 3), strides=(2, 2))(x)
for i in range(0, len(stages)):
stride = (1, 1) if i == 0 else (2, 2)
x = ResNet.residual_module(x,
filters[i],
stride,
chanDim,
red=True,
bnEps=bnEps,
bnMom=bnMom)
for j in range(0, stages[i] - 1):
x = ResNet.residual_module(x,
filters[i], (1, 1),
chanDim,
bnEps=bnEps,
bnMom=bnMom)
x = BatchNormalization(axis=chanDim, epsilon=bnEps, momentum=bnMom)(x)
x = Activation("relu")(x)
x = Conv2D(200, (1, 1), kernel_regularizer=l2(reg))(x)
x = GlobalAveragePooling2D('channels_last')(x)
x = Activation("softmax")(x)
model = Model(inputs, x, name="resnet")
return model
def model1(self):
base_model = InceptionV3(weights='imagenet', include_top=False)
x = base_model.output
x = GlobalAveragePooling2D()(x)
self.model = Model(inputs=base_model.input, outputs=x)
return self.model
def Deeplabv3pa(weights='pascal_voc',
input_tensor=None,
input_shape=(512, 512, 3),
classes=21,
OS=16):
if not (weights in {'pascal_voc', None}):
raise ValueError('The `weights` argument should be either '
'`None` (random initialization) or `pascal_voc` '
'(pre-trained on PASCAL VOC)')
if K.backend() != 'tensorflow':
raise RuntimeError('The Deeplabv3+ model is only available with '
'the TensorFlow backend.')
if OS == 8:
entry_block3_stride = 1
middle_block_rate = 2 # ! Not mentioned in paper, but required
exit_block_rates = (2, 4)
atrous_rates = (12, 24, 36)
else:
entry_block3_stride = 2
middle_block_rate = 1
exit_block_rates = (1, 2)
atrous_rates = (6, 12, 18)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
x = Conv2D(32, (3, 3),
strides=(2, 2),
name='entry_flow_conv1_1',
use_bias=False,
padding='same')(img_input)
x = BatchNormalization(name='entry_flow_conv1_1_BN')(x)
x = Activation('relu')(x)
x = conv2d_same(x, 64, 'entry_flow_conv1_2', kernel_size=3, stride=1)
x = BatchNormalization(name='entry_flow_conv1_2_BN')(x)
x = Activation('relu')(x)
x = xception_block(x, [128, 128, 128],
'entry_flow_block1',
skip_connection_type='conv',
stride=2,
depth_activation=False)
x, skip1 = xception_block(x, [256, 256, 256],
'entry_flow_block2',
skip_connection_type='conv',
stride=2,
depth_activation=False,
return_skip=True)
x = xception_block(x, [728, 728, 728],
'entry_flow_block3',
skip_connection_type='conv',
stride=entry_block3_stride,
depth_activation=False)
for i in range(16):
x = xception_block(x, [728, 728, 728],
'middle_flow_unit_{}'.format(i + 1),
skip_connection_type='sum',
stride=1,
rate=middle_block_rate,
depth_activation=False)
x = xception_block(x, [728, 1024, 1024],
'exit_flow_block1',
skip_connection_type='conv',
stride=1,
rate=exit_block_rates[0],
depth_activation=False)
x = xception_block(x, [1536, 1536, 2048],
'exit_flow_block2',
skip_connection_type='none',
stride=1,
rate=exit_block_rates[1],
depth_activation=True)
# end of feature extractor
# branching for Atrous Spatial Pyramid Pooling
# simple 1x1
b0 = Conv2D(256, (1, 1), padding='same', use_bias=False, name='aspp0')(x)
b0 = BatchNormalization(name='aspp0_BN', epsilon=1e-5)(b0)
b0 = Activation('relu', name='aspp0_activation')(b0)
# rate = 6 (12)
b1 = SepConv_BN(x,
256,
'aspp1',
rate=atrous_rates[0],
depth_activation=True,
epsilon=1e-5)
# rate = 12 (24)
b2 = SepConv_BN(x,
256,
'aspp2',
rate=atrous_rates[1],
depth_activation=True,
epsilon=1e-5)
# rate = 18 (36)
b3 = SepConv_BN(x,
256,
'aspp3',
rate=atrous_rates[2],
depth_activation=True,
epsilon=1e-5)
# Image Feature branch
out_shape = int(np.ceil(input_shape[0] / OS))
b4 = AveragePooling2D(pool_size=(out_shape, out_shape))(x)
b4 = Conv2D(256, (1, 1),
padding='same',
use_bias=False,
name='image_pooling')(b4)
b4 = BatchNormalization(name='image_pooling_BN', epsilon=1e-5)(b4)
b4 = Activation('relu')(b4)
b4 = BilinearUpsampling((out_shape, out_shape), l_name='up1')(b4)
b0_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b0)
b0_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b0_1)
b0_1 = Dropout(0.5)(b0_1)
b0_c = concatenate([b0, b0_1], axis=3)
b0_2 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b0_c)
b0 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b0_2)
# b1
b1_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b1)
b1_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b1_1)
b1_1 = Dropout(0.5)(b1_1)
b1_c = concatenate([b1, b1_1], axis=3)
b1_2 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b1_c)
b1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b1_2)
# b2
b2_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b2)
b2_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b2_1)
b2_1 = Dropout(0.5)(b2_1)
b2_c = concatenate([b2, b2_1], axis=3)
b2_2 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b2_c)
b2 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b2_2)
# b3
b3_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b3)
b3_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b3_1)
b3_1 = Dropout(0.5)(b3_1)
b3_c = concatenate([b3, b3_1], axis=3)
b3_2 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b3_c)
b3 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b3_2)
# b4
b4_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b4)
b4_1 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b4_1)
b4_1 = Dropout(0.5)(b4_1)
b4_c = concatenate([b4, b4_1], axis=3)
b4_2 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b4_c)
b4 = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(b4_2)
Dense0 = Dense(256,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)
Dense1 = Dense(32,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)
Dense2 = Dense(256,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)
b0_1 = Reshape((1, 1, 256))(GlobalAveragePooling2D()(b0))
b0_1 = Dense2(Dense1(Dense0(b0_1)))
b1_1 = Reshape((1, 1, 256))(GlobalAveragePooling2D()(b1))
b1_1 = Dense2(Dense1(Dense0(b1_1)))
b2_1 = Reshape((1, 1, 256))(GlobalAveragePooling2D()(b2))
b2_1 = Dense2(Dense1(Dense0(b2_1)))
b3_1 = Reshape((1, 1, 256))(GlobalAveragePooling2D()(b3))
b3_1 = Dense2(Dense1(Dense0(b3_1)))
b4_1 = Reshape((1, 1, 256))(GlobalAveragePooling2D()(b4))
b4_1 = Dense2(Dense1(Dense0(b4_1)))
x0 = multiply([b0, b0_1])
x1 = multiply([b1, b1_1])
x2 = multiply([b2, b2_1])
x3 = multiply([b3, b3_1])
x4 = multiply([b4, b4_1])
x0 = Reshape((16, 16, 1, 256))(x0)
x1 = Reshape((16, 16, 1, 256))(x1)
x2 = Reshape((16, 16, 1, 256))(x2)
x3 = Reshape((16, 16, 1, 256))(x3)
x4 = Reshape((16, 16, 1, 256))(x4)
x = Concatenate(axis=3)([x0, x1, x2, x3, x4])
x = Conv3D(256, (1, 1, 5),
activation='relu',
use_bias=False,
kernel_initializer='he_normal')(x)
x = Reshape((16, 16, 256))(x)
x = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(x)
x = Conv2D(256, (3, 3),
activation='relu',
padding='same',
kernel_initializer='he_normal')(x)
x = BatchNormalization(name='concat_projection_BN', epsilon=1e-5)(x)
x = Activation('relu')(x)
x = Dropout(0.1)(x)
# DeepLab v.3+ decoder
# Feature projection
x = BilinearUpsampling(output_size=(int(np.ceil(input_shape[0] / 4)),
int(np.ceil(input_shape[1] / 4))),
l_name='up2')(x)
dec_skip1 = Conv2D(48, (1, 1),
padding='same',
use_bias=False,
name='feature_projection0')(skip1)
dec_skip1 = BatchNormalization(name='feature_projection0_BN',
epsilon=1e-5)(dec_skip1)
dec_skip1 = Activation('relu')(dec_skip1)
x = Concatenate()([x, dec_skip1])
x = SepConv_BN(x,
256,
'decoder_conv0',
depth_activation=True,
epsilon=1e-5)
x = SepConv_BN(x,
256,
'decoder_conv1',
depth_activation=True,
epsilon=1e-5)
conv8 = Conv2D(64,
3,
activation='relu',
padding='same',
kernel_initializer='he_normal')(x)
conv8 = Conv2D(64,
3,
activation='relu',
padding='same',
kernel_initializer='he_normal')(conv8)
conv9 = Conv2D(2,
3,
activation='relu',
padding='same',
kernel_initializer='he_normal')(conv8)
x = Conv2D(1, 1, activation='sigmoid')(conv9)
x1 = BilinearUpsampling(output_size=(input_shape[0], input_shape[1]),
l_name='x1')(x)
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
model = Model(inputs, x1, name='deeplabv3+')
# load weights
if weights == 'pascal_voc':
weights_path = get_file(
'deeplabv3_weights_tf_dim_ordering_tf_kernels.h5',
TF_WEIGHTS_PATH,
cache_subdir='models')
model.load_weights(weights_path, by_name=True)
model.compile(optimizer=Adam(lr=1e-4),
loss='binary_crossentropy',
metrics=['accuracy'])
return model
def MobileNet(input_shape=None,
alpha=1.0,
depth_multiplier=1,
dropout=1e-3,
include_top=True,
weights=None,
pooling=None,
classes=10):
"""Instantiates the MobileNet architecture.
Note that only TensorFlow is supported for now,
therefore it only works with the data format
`image_data_format='channels_last'` in your Keras config
at `~/.keras/keras.json`.
To load a MobileNet model via `load_model`, import the custom
objects `relu6` and `DepthwiseConv2D` and pass them to the
`custom_objects` parameter.
E.g.
model = load_model('mobilenet.h5', custom_objects={
'relu6': mobilenet.relu6,
'DepthwiseConv2D': mobilenet.DepthwiseConv2D})
# Arguments
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or (3, 224, 224) (with `channels_first` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 32.
E.g. `(200, 200, 3)` would be one valid value.
alpha: controls the width of the network.
- If `alpha` < 1.0, proportionally decreases the number
of filters in each layer.
- If `alpha` > 1.0, proportionally increases the number
of filters in each layer.
- If `alpha` = 1, default number of filters from the paper
are used at each layer.
depth_multiplier: depth multiplier for depthwise convolution
(also called the resolution multiplier)
dropout: dropout rate
include_top: whether to include the fully-connected
layer at the top of the network.
weights: `None` (random initialization) or
`imagenet` (ImageNet weights)
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model
will be the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a
2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
RuntimeError: If attempting to run this model with a
backend that does not support separable convolutions.
"""
if K.image_data_format() == 'channels_last':
row_axis = 0
else:
row_axis = 1
rows = input_shape[row_axis]
input_tensor = Input(shape=input_shape)
x = _conv_block(input_tensor, 32, alpha, strides=(2, 2))
x = _depthwise_conv_block(x, 64, alpha, depth_multiplier, block_id=1)
x = _depthwise_conv_block(x,
128,
alpha,
depth_multiplier,
strides=(2, 2),
block_id=2)
x = _depthwise_conv_block(x, 128, alpha, depth_multiplier, block_id=3)
x = _depthwise_conv_block(x,
256,
alpha,
depth_multiplier,
strides=(2, 2),
block_id=4)
x = _depthwise_conv_block(x, 256, alpha, depth_multiplier, block_id=5)
x = _depthwise_conv_block(x,
512,
alpha,
depth_multiplier,
strides=(2, 2),
block_id=6)
x = _depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=7)
x = _depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=8)
x = _depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=9)
x = _depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=10)
x = _depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=11)
x = _depthwise_conv_block(x,
1024,
alpha,
depth_multiplier,
strides=(2, 2),
block_id=12)
x = _depthwise_conv_block(x, 1024, alpha, depth_multiplier, block_id=13)
if include_top:
shape = (1, 1, int(1024 * alpha))
x = GlobalAveragePooling2D()(x)
x = Reshape(shape, name='reshape_1')(x)
x = Dropout(dropout, name='dropout')(x)
x = Conv2D(classes, (1, 1), padding='same', name='conv_preds')(x)
x = Activation('softmax', name='act_softmax')(x)
x = Reshape((classes, ), name='reshape_2')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# Create model.
model = Model(input_tensor, x, name='mobilenet_%0.2f_%s' % (alpha, rows))
return model
elif model_choice == 'conv9':
model_9conv(train_x, train_y, test_x, test_y)
else:
# finetuned model - test set accuracy on SVHN: ~94%
# based on the sample finetuning code provided by keras at
# https://keras.io/applications/
base_model = VGG16(weights='imagenet',
include_top=False,
input_tensor=Input(shape=(128, 128, 3)))
print(base_model.summary())
new_top = base_model.output
new_top = GlobalAveragePooling2D()(new_top)
new_top = Dense(512, activation='relu')(new_top)
new_top = Dropout(0.5)(new_top)
new_top = Dense(10, activation='softmax')(new_top)
model = Model(inputs=base_model.input, outputs=new_top)
# freeze vgg16 layers to train new output layers
for layer in base_model.layers:
layer.trainable = False
model.compile(optimizer=SGD(learning_rate=1e-4, momentum=0.9),
loss='categorical_crossentropy',
metrics=['accuracy'])
print('Training new output layers...')
def xnet(
classes=14,
input_shape=(224, 224, 1),
):
X_input = Input(input_shape)
layer = 0
layer = layer + 1
X = ZeroPadding2D((3, 3))(X_input)
X = BatchNormalization(axis=3, name=bn + str(layer))(X)
X = Activation('relu')(X)
X = Conv2D(2 * k, (7, 7),
strides=(2, 2),
name=conv + str(layer),
kernel_initializer=glorot_uniform(seed=0))(X)
X = ZeroPadding2D((1, 1))(X)
X = MaxPooling2D((3, 3), strides=(2, 2))(X)
#Dense Block = 1
layer = layer + 1
X = bottleneck_composite(X, layer)
l = []
l.append(X)
for i in range(0, 5):
layer = layer + 2
X = bottleneck_composite(l, layer)
l.append(X)
# Transition layer = 1
layer = layer + 2
X = BatchNormalization(axis=3, name=bn + str(layer))(X)
X = Activation('relu')(X)
X = Conv2D(k, (1, 1),
strides=(1, 1),
padding='same',
name=conv + str(layer),
kernel_initializer=glorot_uniform(seed=0))(X)
X = ZeroPadding2D((1, 1))(X)
X = AveragePooling2D((2, 2), strides=(2, 2))(X)
#Dense Block = 2
layer = layer + 1
X = bottleneck_composite(X, layer)
l = []
l.append(X)
for i in range(0, 11):
layer = layer + 2
X = bottleneck_composite(l, layer)
l.append(X)
# Transition layer = 2
layer = layer + 2
X = BatchNormalization(axis=3, name=bn + str(layer))(X)
X = Activation('relu')(X)
X = Conv2D(k, (1, 1),
strides=(1, 1),
padding='same',
name=conv + str(layer),
kernel_initializer=glorot_uniform(seed=0))(X)
X = ZeroPadding2D((1, 1))(X)
X = AveragePooling2D((2, 2), strides=(2, 2))(X)
#Dense Block = 3
layer = layer + 1
X = bottleneck_composite(X, layer)
l = []
l.append(X)
for i in range(0, 31):
layer = layer + 2
X = bottleneck_composite(l, layer)
l.append(X)
# Transition layer = 3
layer = layer + 2
X = BatchNormalization(axis=3, name=bn + str(layer))(X)
X = Activation('relu')(X)
X = Conv2D(k, (1, 1),
strides=(1, 1),
padding='same',
name=conv + str(layer),
kernel_initializer=glorot_uniform(seed=0))(X)
X = ZeroPadding2D((1, 1))(X)
X = AveragePooling2D((2, 2), strides=(2, 2))(X)
#Dense Block = 4
layer = layer + 1
X = bottleneck_composite(X, layer)
l = []
l.append(X)
for i in range(0, 31):
layer = layer + 2
X = bottleneck_composite(l, layer)
l.append(X)
layer = layer + 2
X = GlobalAveragePooling2D()(X)
# fully connected layer
#X = Flatten()(X)
X = Dense(classes,
activation='softmax',
name=fc + str(layer),
kernel_initializer=glorot_uniform(seed=0))(X)
model = Model(inputs=X_input, outputs=X, name="DenseNet121")
return model
def VGG16(include_top=True,
weights='imagenet',
input_tensor=None,
input_shape=None,
pooling=None,
classes=1000):
"""Instantiates the VGG16 architecture.
Optionally loads weights pre-trained
on ImageNet. Note that when using TensorFlow,
for best performance you should set
`image_data_format="channels_last"` in your Keras config
at ~/.keras/keras.json.
The model and the weights are compatible with both
TensorFlow and Theano. The data format
convention used by the model is the one
specified in your Keras config file.
# Arguments
include_top: whether to include the 3 fully-connected
layers at the top of the network.
weights: one of `None` (random initialization)
or "imagenet" (pre-training on ImageNet).
input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
to use as image input for the model.
input_shape: optional shape tuple, only to be specified
if `include_top` is False (otherwise the input shape
has to be `(224, 224, 3)` (with `channels_last` data format)
or `(3, 224, 244)` (with `channels_first` data format).
It should have exactly 3 inputs channels,
and width and height should be no smaller than 48.
E.g. `(200, 200, 3)` would be one valid value.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model will be
the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a 2D tensor.
- `max` means that global max pooling will
be applied.
classes: optional number of classes to classify images
into, only to be specified if `include_top` is True, and
if no `weights` argument is specified.
# Returns
A Keras model instance.
# Raises
ValueError: in case of invalid argument for `weights`,
or invalid input shape.
"""
if weights not in {'imagenet', None}:
raise ValueError('The `weights` argument should be either '
'`None` (random initialization) or `imagenet` '
'(pre-training on ImageNet).')
if weights == 'imagenet' and include_top and classes != 1000:
raise ValueError('If using `weights` as imagenet with `include_top`'
' as true, `classes` should be 1000')
# Determine proper input shape
input_shape = _obtain_input_shape(input_shape,
default_size=224,
min_size=48,
data_format=K.image_data_format(),
require_flatten=False)
if input_tensor is None:
img_input = Input(shape=input_shape)
else:
if not K.is_keras_tensor(input_tensor):
img_input = Input(tensor=input_tensor, shape=input_shape)
else:
img_input = input_tensor
# Block 1
x = Conv2D(64, (3, 3),
activation='relu',
padding='same',
name='block1_conv1')(img_input)
x = Conv2D(64, (3, 3),
activation='relu',
padding='same',
name='block1_conv2')(x)
x = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)
# Block 2
x = Conv2D(128, (3, 3),
activation='relu',
padding='same',
name='block2_conv1')(x)
x = Conv2D(128, (3, 3),
activation='relu',
padding='same',
name='block2_conv2')(x)
x = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)
# Block 3
x = Conv2D(256, (3, 3),
activation='relu',
padding='same',
name='block3_conv1')(x)
x = Conv2D(256, (3, 3),
activation='relu',
padding='same',
name='block3_conv2')(x)
x = Conv2D(256, (3, 3),
activation='relu',
padding='same',
name='block3_conv3')(x)
x = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)
# Block 4
x = Conv2D(512, (3, 3),
activation='relu',
padding='same',
name='block4_conv1')(x)
x = Conv2D(512, (3, 3),
activation='relu',
padding='same',
name='block4_conv2')(x)
x = Conv2D(512, (3, 3),
activation='relu',
padding='same',
name='block4_conv3')(x)
x = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)
# Block 5
x = Conv2D(512, (3, 3),
activation='relu',
padding='same',
name='block5_conv1')(x)
x = Conv2D(512, (3, 3),
activation='relu',
padding='same',
name='block5_conv2')(x)
x = Conv2D(512, (3, 3),
activation='relu',
padding='same',
name='block5_conv3')(x)
x = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)
if include_top:
# Classification block
x = Flatten(name='flatten')(x)
x = Dense(4096, activation='relu', name='fc1')(x)
x = Dense(4096, activation='relu', name='fc2')(x)
x = Dense(classes, activation='softmax', name='predictions')(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
# Ensure that the model takes into account
# any potential predecessors of `input_tensor`.
if input_tensor is not None:
inputs = get_source_inputs(input_tensor)
else:
inputs = img_input
# Create model.
model = Model(inputs, x, name='vgg16')
# load weights
if weights == 'imagenet':
if include_top:
weights_path = get_file(
'vgg16_weights_tf_dim_ordering_tf_kernels.h5',
WEIGHTS_PATH,
)
else:
weights_path = get_file(
'vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5',
WEIGHTS_PATH_NO_TOP,
cache_subdir='models')
model.load_weights(weights_path)
if K.backend() == 'theano':
layer_utils.convert_all_kernels_in_model(model)
if K.image_data_format() == 'channels_first':
if include_top:
maxpool = model.get_layer(name='block5_pool')
shape = maxpool.output_shape[1:]
dense = model.get_layer(name='fc1')
layer_utils.convert_dense_weights_data_format(
dense, shape, 'channels_first')
if K.backend() == 'tensorflow':
warnings.warn('You are using the TensorFlow backend, yet you '
'are using the Theano '
'image data format convention '
'(`image_data_format="channels_first"`). '
'For best performance, set '
'`image_data_format="channels_last"` in '
'your Keras config '
'at ~/.keras/keras.json.')
model.layers.pop() # Get rid of the classification layer
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes = []
return model
model.add(Activation('relu'))
model.add(AveragePooling2D(pool_size=(2, 2), strides=1, padding='same'))
model.add(Dropout(0.5))
model.add(Conv2D(filters=128, kernel_size=(3, 3), padding='same'))
model.add(BatchNormalization())
model.add(Conv2D(filters=128, kernel_size=(3, 3), padding='same'))
model.add(BatchNormalization())
model.add(Activation('relu'))
model.add(AveragePooling2D(pool_size=(2, 2), padding='same'))
model.add(Dropout(0.5))
model.add(Conv2D(filters=256, kernel_size=(3, 3), padding='same'))
model.add(BatchNormalization())
model.add(Conv2D(filters=nb_classes, kernel_size=(3, 3), padding='same'))
model.add(GlobalAveragePooling2D())
model.add(Activation('softmax', name='predictions'))
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
#{0:'neutral', 1:'anger', 2:'contempt', 3:'disgust', 4:'fear', 5:'happy', 6:'sadness', 7:'surprise'}
"""
class_weights = {0: 0.000761615,
1: 0.001290323,
2: 0.005376344,
3: 0.001451379,
4: 0.002398082,
5: 0.000970874,
6: 0.002380952,
validation_data=(X_test, y_test))
print('Training time: %s' % (t - time.time()))
(loss, accuracy) = custom_resnet_model.evaluate(X_test,
y_test,
batch_size=10,
verbose=1)
print("[INFO] loss={:.4f}, accuracy: {:.4f}%".format(loss, accuracy * 100))
###########################################################################################################################
# Fine tune the resnet 50
#image_input = Input(shape=(224, 224, 3))
model = ResNet50(weights='imagenet', include_top=False)
model.summary()
last_layer = model.output
# add a global spatial average pooling layer
x = GlobalAveragePooling2D()(last_layer)
# add fully-connected & dropout layers
x = Dense(512, activation='relu', name='fc-1')(x)
x = Dropout(0.5)(x)
x = Dense(256, activation='relu', name='fc-2')(x)
x = Dropout(0.5)(x)
# a softmax layer for 4 classes
out = Dense(num_classes, activation='softmax', name='output_layer')(x)
# this is the model we will train
custom_resnet_model2 = Model(inputs=model.input, outputs=out)
custom_resnet_model2.summary()
for layer in custom_resnet_model2.layers[:-6]:
layer.trainable = False
custom_resnet_model2.layers[-1].trainable
def _main(args):
config_path = os.path.expanduser(args.config_path)
weights_path = os.path.expanduser(args.weights_path)
assert config_path.endswith('.cfg'), '{} is not a .cfg file'.format(
config_path)
assert weights_path.endswith(
'.weights'), '{} is not a .weights file'.format(weights_path)
output_path = os.path.expanduser(args.output_path)
assert output_path.endswith(
'.h5'), 'output path {} is not a .h5 file'.format(output_path)
output_root = os.path.splitext(output_path)[0]
# Load weights and config.
print('Loading weights.')
weights_file = open(weights_path, 'rb')
weights_header = np.ndarray(shape=(4, ),
dtype='int32',
buffer=weights_file.read(16))
print('Weights Header: ', weights_header)
# TODO: Check transpose flag when implementing fully connected layers.
# transpose = (weight_header[0] > 1000) or (weight_header[1] > 1000)
print('Parsing Darknet config.')
unique_config_file = unique_config_sections(config_path)
cfg_parser = configparser.ConfigParser()
cfg_parser.read_file(unique_config_file)
print('Creating Keras model.')
if args.fully_convolutional:
image_height, image_width = None, None
else:
image_height = int(cfg_parser['net_0']['height'])
image_width = int(cfg_parser['net_0']['width'])
prev_layer = Input(shape=(image_height, image_width, 3))
all_layers = [prev_layer]
weight_decay = float(cfg_parser['net_0']['decay']
) if 'net_0' in cfg_parser.sections() else 5e-4
count = 0
for section in cfg_parser.sections():
print('Parsing section {}'.format(section))
if section.startswith('convolutional'):
filters = int(cfg_parser[section]['filters'])
size = int(cfg_parser[section]['size'])
stride = int(cfg_parser[section]['stride'])
pad = int(cfg_parser[section]['pad'])
activation = cfg_parser[section]['activation']
batch_normalize = 'batch_normalize' in cfg_parser[section]
# padding='same' is equivalent to Darknet pad=1
padding = 'same' if pad == 1 else 'valid'
# Setting weights.
# Darknet serializes convolutional weights as:
# [bias/beta, [gamma, mean, variance], conv_weights]
prev_layer_shape = K.int_shape(prev_layer)
# TODO: This assumes channel last dim_ordering.
weights_shape = (size, size, prev_layer_shape[-1], filters)
darknet_w_shape = (filters, weights_shape[2], size, size)
weights_size = np.product(weights_shape)
print('conv2d', 'bn' if batch_normalize else ' ', activation,
weights_shape)
conv_bias = np.ndarray(shape=(filters, ),
dtype='float32',
buffer=weights_file.read(filters * 4))
count += filters
if batch_normalize:
bn_weights = np.ndarray(shape=(3, filters),
dtype='float32',
buffer=weights_file.read(filters * 12))
count += 3 * filters
# TODO: Keras BatchNormalization mistakenly refers to var
# as std.
bn_weight_list = [
bn_weights[0], # scale gamma
conv_bias, # shift beta
bn_weights[1], # running mean
bn_weights[2] # running var
]
conv_weights = np.ndarray(shape=darknet_w_shape,
dtype='float32',
buffer=weights_file.read(weights_size *
4))
count += weights_size
# DarkNet conv_weights are serialized Caffe-style:
# (out_dim, in_dim, height, width)
# We would like to set these to Tensorflow order:
# (height, width, in_dim, out_dim)
# TODO: Add check for Theano dim ordering.
conv_weights = np.transpose(conv_weights, [2, 3, 1, 0])
conv_weights = [conv_weights] if batch_normalize else [
conv_weights, conv_bias
]
# Handle activation.
act_fn = None
if activation == 'leaky':
pass # Add advanced activation later.
elif activation != 'linear':
raise ValueError(
'Unknown activation function `{}` in section {}'.format(
activation, section))
# Create Conv2D layer
conv_layer = (Conv2D(filters, (size, size),
strides=(stride, stride),
kernel_regularizer=l2(weight_decay),
use_bias=not batch_normalize,
weights=conv_weights,
activation=act_fn,
padding=padding))(prev_layer)
if batch_normalize:
conv_layer = (BatchNormalization(
weights=bn_weight_list))(conv_layer)
prev_layer = conv_layer
if activation == 'linear':
all_layers.append(prev_layer)
elif activation == 'leaky':
act_layer = LeakyReLU(alpha=0.1)(prev_layer)
prev_layer = act_layer
all_layers.append(act_layer)
elif section.startswith('maxpool'):
size = int(cfg_parser[section]['size'])
stride = int(cfg_parser[section]['stride'])
all_layers.append(
MaxPooling2D(padding='same',
pool_size=(size, size),
strides=(stride, stride))(prev_layer))
prev_layer = all_layers[-1]
elif section.startswith('avgpool'):
if cfg_parser.items(section) != []:
raise ValueError('{} with params unsupported.'.format(section))
all_layers.append(GlobalAveragePooling2D()(prev_layer))
prev_layer = all_layers[-1]
elif section.startswith('route'):
ids = [int(i) for i in cfg_parser[section]['layers'].split(',')]
layers = [all_layers[i] for i in ids]
if len(layers) > 1:
print('Concatenating route layers:', layers)
concatenate_layer = concatenate(layers)
all_layers.append(concatenate_layer)
prev_layer = concatenate_layer
else:
skip_layer = layers[0] # only one layer to route
all_layers.append(skip_layer)
prev_layer = skip_layer
elif section.startswith('reorg'):
block_size = int(cfg_parser[section]['stride'])
assert block_size == 2, 'Only reorg with stride 2 supported.'
all_layers.append(
Lambda(space_to_depth_x2,
output_shape=space_to_depth_x2_output_shape,
name='space_to_depth_x2')(prev_layer))
prev_layer = all_layers[-1]
elif section.startswith('region'):
with open('{}_anchors.txt'.format(output_root), 'w') as f:
print(cfg_parser[section]['anchors'], file=f)
elif (section.startswith('net') or section.startswith('cost')
or section.startswith('softmax')):
pass # Configs not currently handled during model definition.
else:
raise ValueError(
'Unsupported section header type: {}'.format(section))
# Create and save model.
model = Model(inputs=all_layers[0], outputs=all_layers[-1])
print(model.summary())
model.save('{}'.format(output_path))
print('Saved Keras model to {}'.format(output_path))
# Check to see if all weights have been read.
remaining_weights = len(weights_file.read()) / 4
weights_file.close()
print('Read {} of {} from Darknet weights.'.format(
count, count + remaining_weights))
if remaining_weights > 0:
print('Warning: {} unused weights'.format(remaining_weights))
if args.plot_model:
plot(model, to_file='{}.png'.format(output_root), show_shapes=True)
print('Saved model plot to {}.png'.format(output_root))
'/home/liuk/dl/pytorch/data/newdata/val',
target_size=(256, 256),
#target_size=(299,299),
batch_size=16,
class_mode='categorical',
shuffle=True)
# 构建基础模型
#base_model = InceptionV3(weights='imagenet',include_top=True)
base_model = InceptionV3(
weights='imagenet', include_top=False
) #include_top，如果是True，输出是1000个节点的全连接层。如果是False，会去掉顶层，输出一个8 * 8 * 2048的张量
# 增加新的输出层
x = base_model.output
x = GlobalAveragePooling2D()(
x) # GlobalAveragePooling2D 将 MxNxC 的张量转换成1xC 张量，C是通道数
x = Dense(1024, activation='relu')(x)
predictions = Dense(4, activation='softmax')(x)
model = Model(inputs=base_model.input, outputs=predictions)
# plot_model(model,'tlmodel.png')
plot_model(model, 'output/tlmodel.png')
'''
这里的base_model和model里面的iv3都指向同一个地址
'''
def setup_to_transfer_learning(model, base_model): #base_model
for layer in base_model.layers:
layer.trainable = False
model.compile(optimizer=RMSprop(1e-3),
def SqueezeNet(nb_classes, inputs=(224, 224, 3)):
""" Keras Implementation of SqueezeNet(arXiv 1602.07360)
@param nb_classes: total number of final categories
Arguments:
inputs -- shape of the input images (channel, cols, rows)
"""
input_img = Input(shape=inputs)
conv1 = Convolution2D(96, (7, 7),
activation='relu',
kernel_initializer='glorot_uniform',
strides=(2, 2),
padding='same',
name='conv1')(input_img)
maxpool1 = MaxPooling2D(pool_size=(3, 3), strides=(2, 2),
name='maxpool1')(conv1)
fire2_squeeze = Convolution2D(16, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire2_squeeze')(maxpool1)
fire2_expand1 = Convolution2D(64, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire2_expand1')(fire2_squeeze)
fire2_expand2 = Convolution2D(64, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire2_expand2')(fire2_squeeze)
merge2 = Concatenate(axis=1)([fire2_expand1, fire2_expand2])
fire3_squeeze = Convolution2D(16, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire3_squeeze')(merge2)
fire3_expand1 = Convolution2D(64, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire3_expand1')(fire3_squeeze)
fire3_expand2 = Convolution2D(64, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire3_expand2')(fire3_squeeze)
merge3 = Concatenate(axis=1)([fire3_expand1, fire3_expand2])
fire4_squeeze = Convolution2D(32, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire4_squeeze')(merge3)
fire4_expand1 = Convolution2D(128, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire4_expand1')(fire4_squeeze)
fire4_expand2 = Convolution2D(128, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire4_expand2')(fire4_squeeze)
merge4 = Concatenate(axis=1)([fire4_expand1, fire4_expand2])
maxpool4 = MaxPooling2D(pool_size=(3, 3), strides=(2, 2),
name='maxpool4')(merge4)
fire5_squeeze = Convolution2D(32, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire5_squeeze')(maxpool4)
fire5_expand1 = Convolution2D(128, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire5_expand1')(fire5_squeeze)
fire5_expand2 = Convolution2D(128, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire5_expand2')(fire5_squeeze)
merge5 = Concatenate(axis=1)([fire5_expand1, fire5_expand2])
fire6_squeeze = Convolution2D(48, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire6_squeeze')(merge5)
fire6_expand1 = Convolution2D(192, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire6_expand1')(fire6_squeeze)
fire6_expand2 = Convolution2D(192, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire6_expand2')(fire6_squeeze)
merge6 = Concatenate(axis=1)([fire6_expand1, fire6_expand2])
fire7_squeeze = Convolution2D(48, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire7_squeeze')(merge6)
fire7_expand1 = Convolution2D(192, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire7_expand1')(fire7_squeeze)
fire7_expand2 = Convolution2D(192, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire7_expand2')(fire7_squeeze)
merge7 = Concatenate(axis=1)([fire7_expand1, fire7_expand2])
fire8_squeeze = Convolution2D(64, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire8_squeeze')(merge7)
fire8_expand1 = Convolution2D(256, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire8_expand1')(fire8_squeeze)
fire8_expand2 = Convolution2D(256, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire8_expand2')(fire8_squeeze)
merge8 = Concatenate(axis=1)([fire8_expand1, fire8_expand2])
maxpool8 = MaxPooling2D(pool_size=(3, 3), strides=(2, 2),
name='maxpool8')(merge8)
fire9_squeeze = Convolution2D(64, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire9_squeeze')(maxpool8)
fire9_expand1 = Convolution2D(256, (1, 1),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire9_expand1')(fire9_squeeze)
fire9_expand2 = Convolution2D(256, (3, 3),
activation='relu',
kernel_initializer='glorot_uniform',
padding='same',
name='fire9_expand2')(fire9_squeeze)
merge9 = Concatenate(axis=1)([fire9_expand1, fire9_expand2])
fire9_dropout = Dropout(0.5, name='fire9_dropout')(merge9)
conv10 = Convolution2D(nb_classes, (1, 1),
kernel_initializer='glorot_uniform',
padding='valid',
name='conv10')(fire9_dropout)
global_avgpool10 = GlobalAveragePooling2D()(conv10)
softmax = Activation("softmax", name='softmax')(global_avgpool10)
return Model(inputs=input_img, outputs=softmax)
def get_model(embed, num_conti):
def block_wrap(x, filters):
x = Conv2D(filters, kernel_size=3, padding='same')(x)
x = BatchNormalization()(x)
x = Activation('relu')(x)
return x
def root_mean_squared_error(y_true, y_pred):
return K.sqrt(K.mean(K.square(y_true - y_pred)))
region = Input(shape=[1], name="region")
city = Input(shape=[1], name="city")
pcn = Input(shape=[1], name="parent_category_name")
cn = Input(shape=[1], name="category_name")
para1 = Input(shape=[1], name="param_1")
para2 = Input(shape=[1], name="param_2")
para3 = Input(shape=[1], name="param_3")
# act = Input(shape=[1], name="activation_date")
ut = Input(shape=[1], name="user_type")
img_top = Input(shape=[1], name="image_top_1")
title = Input(shape=[title_len], name="title")
desc = Input(shape=[desc_len], name="desc")
img = Input(shape=[img_size, img_size, 3], name='image')
conti = Input(shape=[num_conti], name='conti')
"""
region 28
city 1752
parent_category_name 9
category_name 47
param_1 372
param_2 278
param_3 1277
title 1022203
description 1793973
activation_date 30
user_type 3
image 1856666
image_top_1 3064
param_combined 2402
"""
emb_region = Embedding(28, 8)(region)
emb_city = Embedding(1752, 16)(city)
emb_pcn = Embedding(9, 3)(pcn)
emb_cn = Embedding(47, 8)(cn)
emb_para1 = Embedding(372, 16)(para1)
emb_para2 = Embedding(278, 16)(para2)
emb_para3 = Embedding(1277, 16)(para3)
# emb_act = Embedding(30,8)(act)
emb_img_top = Embedding(3064, 32)(img_top)
emb_ut = Embedding(3, 3, weights=[np.eye(3, 3)], trainable=False)(ut)
num_word = len(embed['title'])
emb_title = Embedding(num_word,
300,
weights=[embed['title']],
trainable=False)(title)
num_word = len(embed['desc'])
emb_desc = Embedding(num_word,
300,
weights=[embed['desc']],
trainable=False)(desc)
conv = block_wrap(img, 32)
conv = MaxPool2D(padding='same')(conv)
conv = block_wrap(conv, 64)
conv = MaxPool2D(padding='same')(conv)
conv = block_wrap(conv, 128)
conv = MaxPool2D(padding='same')(conv)
conv = GlobalAveragePooling2D()(conv)
title_gru = Bidirectional(CuDNNGRU(64, return_sequences=True),
merge_mode='sum')(emb_title)
title_gru = Bidirectional(CuDNNGRU(32, return_sequences=True),
merge_mode='sum')(title_gru)
title_gru1 = GlobalMaxPooling1D()(title_gru)
title_gru2 = GlobalAveragePooling1D()(title_gru)
desc_gru = Bidirectional(CuDNNGRU(64, return_sequences=True),
merge_mode='sum')(emb_desc)
desc_gru = Bidirectional(CuDNNGRU(64, return_sequences=True),
merge_mode='sum')(desc_gru)
desc_gru1 = GlobalMaxPooling1D()(desc_gru)
desc_gru2 = GlobalAveragePooling1D()(desc_gru)
fc = concatenate([
Flatten()(emb_region),
Flatten()(emb_city),
Flatten()(emb_pcn),
Flatten()(emb_cn),
Flatten()(emb_para1),
Flatten()(emb_para2),
Flatten()(emb_para3),
# Flatten()(emb_act),
Flatten()(emb_img_top),
Flatten()(emb_ut),
conti,
title_gru1,
title_gru2,
desc_gru1,
desc_gru2,
conv
])
fc = Dense(256, activation='relu')(fc)
fc = Dropout(0.5)(fc)
fc = Dense(1, activation='sigmoid', name='output')(fc)
model = Model(
[
region,
city,
pcn,
cn,
para1,
para2,
para3,
# act,
ut,
img_top,
title,
desc,
img,
conti,
],
output=fc)
model.compile(optimizer=Nadam(),
loss="mean_squared_error",
metrics=[root_mean_squared_error])
return model
import tflearn.datasets.oxflower17 as oxflower17
from sklearn.model_selection import train_test_split
X, Y = oxflower17.load_data(one_hot=True, resize_pics=(224, 224))
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2)
# Step 1: Create the base pre-trained model
input_tensor = Input(shape=(224, 224, 3))
base_model = InceptionV3(input_tensor=input_tensor,
weights='imagenet',
include_top=False)
# Step 2: Create a new model with dense and softamx layers
x = base_model.output
x = GlobalAveragePooling2D()(x)
x = Dense(1024, activation='relu')(x)
predictions = Dense(17, activation='softmax')(x)
model = Model(inputs=base_model.input, outputs=predictions)
# Step 3: Freeze all pre-trained layers and train the top layers with new dataaset
for layer in base_model.layers:
layer.trainable = False
model.compile(optimizer='adadelta',
loss='categorical_crossentropy',
metrics=['accuracy'])
model.fit(X_train, Y_train, batch_size=50, epochs=10)
# Step 4: Unfreeze some pre-trained layers and train with new dataset
for layer in model.layers[:5]:
layer.trainable = False
def __create_dense_net(nb_classes, img_input, include_top, depth=40, nb_dense_block=3, growth_rate=12, nb_filter=-1,
nb_layers_per_block=-1, bottleneck=False, reduction=0.0, dropout_rate=None, weight_decay=1e-4,
subsample_initial_block=False, pooling=None, activation='softmax', transition_pooling='avg'):
''' Build the DenseNet model
# Arguments
nb_classes: number of classes
img_input: tuple of shape (channels, rows, columns) or (rows, columns, channels)
include_top: flag to include the final Dense layer
depth: number or layers
nb_dense_block: number of dense blocks to add to end (generally = 3)
growth_rate: number of filters to add per dense block
nb_filter: initial number of filters. Default -1 indicates initial number of filters is 2 * growth_rate
nb_layers_per_block: number of layers in each dense block.
Can be a -1, positive integer or a list.
If -1, calculates nb_layer_per_block from the depth of the network.
If positive integer, a set number of layers per dense block.
If list, nb_layer is used as provided. Note that list size must
be (nb_dense_block + 1)
bottleneck: add bottleneck blocks
reduction: reduction factor of transition blocks. Note : reduction value is inverted to compute compression
dropout_rate: dropout rate
weight_decay: weight decay rate
subsample_initial_block: Changes model type to suit different datasets.
Should be set to True for ImageNet, and False for CIFAR datasets.
When set to True, the initial convolution will be strided and
adds a MaxPooling2D before the initial dense block.
pooling: Optional pooling mode for feature extraction
when `include_top` is `False`.
- `None` means that the output of the model
will be the 4D tensor output of the
last convolutional layer.
- `avg` means that global average pooling
will be applied to the output of the
last convolutional layer, and thus
the output of the model will be a
2D tensor.
- `max` means that global max pooling will
be applied.
activation: Type of activation at the top layer. Can be one of 'softmax' or 'sigmoid'.
Note that if sigmoid is used, classes must be 1.
transition_pooling: `avg` for avg pooling (default), `max` for max pooling,
None for no pooling during scale transition blocks. Please note that this
default differs from the DenseNetFCN paper in accordance with the DenseNet
paper.
# Returns
a keras tensor
# Raises
ValueError: in case of invalid argument for `reduction`
or `nb_dense_block`
'''
with K.name_scope('DenseNet'):
concat_axis = 1 if K.image_data_format() == 'channels_first' else -1
if reduction != 0.0:
if not (reduction <= 1.0 and reduction > 0.0):
raise ValueError('`reduction` value must lie between 0.0 and 1.0')
# layers in each dense block
if type(nb_layers_per_block) is list or type(nb_layers_per_block) is tuple:
nb_layers = list(nb_layers_per_block) # Convert tuple to list
if len(nb_layers) != (nb_dense_block):
raise ValueError('If `nb_dense_block` is a list, its length must match '
'the number of layers provided by `nb_layers`.')
final_nb_layer = nb_layers[-1]
nb_layers = nb_layers[:-1]
else:
if nb_layers_per_block == -1:
assert (depth - 4) % 3 == 0, 'Depth must be 3 N + 4 if nb_layers_per_block == -1'
count = int((depth - 4) / 3)
if bottleneck:
count = count // 2
nb_layers = [count for _ in range(nb_dense_block)]
final_nb_layer = count
else:
final_nb_layer = nb_layers_per_block
nb_layers = [nb_layers_per_block] * nb_dense_block
# compute initial nb_filter if -1, else accept users initial nb_filter
if nb_filter <= 0:
nb_filter = 2 * growth_rate
# compute compression factor
compression = 1.0 - reduction
# Initial convolution
if subsample_initial_block:
initial_kernel = (7, 7)
initial_strides = (2, 2)
else:
initial_kernel = (3, 3)
initial_strides = (1, 1)
x = Conv2D(nb_filter, initial_kernel, kernel_initializer='he_normal', padding='same', name='initial_conv2D',
strides=initial_strides, use_bias=False, kernel_regularizer=l2(weight_decay))(img_input)
if subsample_initial_block:
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name='initial_bn')(x)
x = Activation('relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2), padding='same')(x)
# Add dense blocks
for block_idx in range(nb_dense_block - 1):
x, nb_filter = __dense_block(x, nb_layers[block_idx], nb_filter, growth_rate, bottleneck=bottleneck,
dropout_rate=dropout_rate, weight_decay=weight_decay,
block_prefix='dense_%i' % block_idx)
# add transition_block
x = __transition_block(x, nb_filter, compression=compression, weight_decay=weight_decay,
block_prefix='tr_%i' % block_idx, transition_pooling=transition_pooling)
nb_filter = int(nb_filter * compression)
# The last dense_block does not have a transition_block
x, nb_filter = __dense_block(x, final_nb_layer, nb_filter, growth_rate, bottleneck=bottleneck,
dropout_rate=dropout_rate, weight_decay=weight_decay,
block_prefix='dense_%i' % (nb_dense_block - 1))
x = BatchNormalization(axis=concat_axis, epsilon=1.1e-5, name='final_bn')(x)
x = Activation('relu')(x)
if include_top:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
x = Dense(nb_classes, activation=activation)(x)
else:
if pooling == 'avg':
x = GlobalAveragePooling2D()(x)
elif pooling == 'max':
x = GlobalMaxPooling2D()(x)
return x
