def get_seq_model():
"""Define three channel input shape depending on image data format."""
if K.image_data_format() == 'channels_first':
input_shape = (3, img_width, img_height)
else:
input_shape = (img_width, img_height, 3)
# Initialize CNN by creating a sequential model.
model = Sequential()
model.add(Conv2D(32, (3, 3), input_shape=input_shape))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(32, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(64))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(2))
model.add(Activation('sigmoid'))
model.compile(
loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
return model
def squeeze_excite_block(tensor, ratio=16):
init = tensor
channel_axis = 1 if K.image_data_format() == "channels_first" else -1
filters = init.shape[channel_axis]
se_shape = (1, 1, filters)
se = GlobalAveragePooling2D()(init)
se = Reshape(se_shape)(se)
se = Dense(filters // ratio,
kernel_initializer='he_normal',
use_bias=False)(se)
se = LeakyReLU(alpha=0.1)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
if K.image_data_format() == 'channels_first':
se = Permute((3, 1, 2))(se)
x = multiply([init, se])
return x
def _handle_data_format():
global DIM1_AXIS
global DIM2_AXIS
global DIM3_AXIS
global CHANNEL_AXIS
if K.image_data_format() == 'channels_last':
DIM1_AXIS = 1
DIM2_AXIS = 2
DIM3_AXIS = 3
CHANNEL_AXIS = 4
else:
CHANNEL_AXIS = 1
DIM1_AXIS = 2
DIM2_AXIS = 3
DIM3_AXIS = 4
def identity_block_base(input_tensor,
kernel_size,
filters,
stage,
block,
num_updates,
dropout_rate=0.,
use_variational_layers=False):
"""The identity block is the block that has no conv layer at shortcut.
Arguments:
input_tensor: input tensor
kernel_size: default 3, the kernel size of
middle conv layer at main path
filters: list of integers, the filters of 3 conv layer at main path
stage: integer, current stage label, used for generating layer names
block: 'a','b'..., current block label, used for generating layer names
num_updates: integer, total steps in an epoch (for weighting the loss)
dropout_rate: float, always-on dropout rate.
use_variational_layers: boolean, if true train a variational model
Returns:
x: Output tensor for the block.
"""
filters1, filters2, filters3 = filters
divergence_fn = lambda q, p, ignore: (tfd.kl_divergence(q, p) / num_updates
)
if backend.image_data_format() == 'channels_last':
bn_axis = 3
else:
bn_axis = 1
conv_name_base = 'res' + str(stage) + block + '_branch'
bn_name_base = 'bn' + str(stage) + block + '_branch'
if not use_variational_layers:
first_conv_2d = layers.Conv2D(
filters1, (1, 1),
use_bias=False,
kernel_initializer='he_normal',
kernel_regularizer=regularizers.l2(L2_WEIGHT_DECAY),
name=conv_name_base + '2a')
if dropout_rate > 0.:
x = layers.Dropout(dropout_rate)(input_tensor, training=True)
x = first_conv_2d(x)
else:
x = first_conv_2d(input_tensor)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name=bn_name_base + '2a')(x)
x = layers.Activation('relu')(x)
if dropout_rate > 0.:
x = layers.Dropout(dropout_rate)(x, training=True)
x = layers.Conv2D(filters2,
kernel_size,
use_bias=False,
padding='same',
kernel_initializer='he_normal',
kernel_regularizer=regularizers.l2(L2_WEIGHT_DECAY),
name=conv_name_base + '2b')(x)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name=bn_name_base + '2b')(x)
x = layers.Activation('relu')(x)
if dropout_rate > 0.:
x = layers.Dropout(dropout_rate)(x, training=True)
x = layers.Conv2D(filters3, (1, 1),
use_bias=False,
kernel_initializer='he_normal',
kernel_regularizer=regularizers.l2(L2_WEIGHT_DECAY),
name=conv_name_base + '2c')(x)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name=bn_name_base + '2c')(x)
else:
x = tfpl.Convolution2DFlipout(
filters1,
kernel_size=(1, 1),
padding='SAME',
name=conv_name_base + '2a',
kernel_divergence_fn=divergence_fn,
)(input_tensor)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name=bn_name_base + '2a')(x)
x = layers.Activation('relu')(x)
x = tfpl.Convolution2DFlipout(
filters2,
kernel_size=kernel_size,
padding='SAME',
activation=None,
name=conv_name_base + '2b',
kernel_divergence_fn=divergence_fn,
)(x)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name=bn_name_base + '2b')(x)
x = layers.Activation('relu')(x)
x = tfpl.Convolution2DFlipout(
filters3,
kernel_size=(1, 1),
padding='SAME',
activation=None,
name=conv_name_base + '2c',
kernel_divergence_fn=divergence_fn,
)(x)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name=bn_name_base + '2c')(x)
x = layers.add([x, input_tensor])
x = layers.Activation('relu')(x)
return x
def ResNet50(method, num_classes, num_updates, dropout_rate):
"""Instantiates the ResNet50 architecture.
Args:
method: `str`, method for accounting for uncertainty. Must be one of
['vanilla', 'll_dropout', 'll_svi', 'dropout', 'svi', 'dropout_nofirst']
num_classes: `int` number of classes for image classification.
num_updates: integer, total steps in an epoch (for weighting the loss)
dropout_rate: Dropout rate for ll_dropout, dropout methods.
Returns:
A Keras model instance.
pylint: disable=invalid-name
"""
# Determine proper input shape
if backend.image_data_format() == 'channels_first':
input_shape = (3, 224, 224)
bn_axis = 1
else:
input_shape = (224, 224, 3)
bn_axis = 3
if (method in ['dropout', 'll_dropout', 'dropout_nofirst'
]) != (dropout_rate > 0.):
raise ValueError(
'Dropout rate should be nonzero iff a dropout method is used.'
'Method is {}, dropout is {}.'.format(method, dropout_rate))
use_variational_layers = method == 'svi'
hidden_layer_dropout = dropout_rate if method in [
'dropout', 'dropout_nofirst'
] else 0.
img_input = layers.Input(shape=input_shape)
x = layers.ZeroPadding2D(padding=(3, 3), name='conv1_pad')(img_input)
if (dropout_rate > 0.) and (method != 'dropout_nofirst'):
x = layers.Dropout(hidden_layer_dropout)(x, training=True)
x = layers.Conv2D(64, (7, 7),
use_bias=False,
strides=(2, 2),
padding='valid',
kernel_initializer='he_normal',
kernel_regularizer=regularizers.l2(L2_WEIGHT_DECAY),
name='conv1')(x)
x = layers.BatchNormalization(axis=bn_axis,
momentum=BATCH_NORM_DECAY,
epsilon=BATCH_NORM_EPSILON,
name='bn_conv1')(x)
x = layers.Activation('relu')(x)
x = layers.ZeroPadding2D(padding=(1, 1), name='pool1_pad')(x)
x = layers.MaxPooling2D((3, 3), strides=(2, 2))(x)
conv_block = functools.partial(
conv_block_base,
num_updates=num_updates,
dropout_rate=hidden_layer_dropout,
use_variational_layers=use_variational_layers)
identity_block = functools.partial(
identity_block_base,
num_updates=num_updates,
dropout_rate=hidden_layer_dropout,
use_variational_layers=use_variational_layers)
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 = layers.GlobalAveragePooling2D(name='avg_pool')(x)
if dropout_rate > 0.:
x = layers.Dropout(dropout_rate)(x, training=True)
if method in ['ll_svi', 'svi']:
x = tfpl.dense_variational_v2.DenseVariational(
units=num_classes,
make_posterior_fn=posterior_mean_field,
make_prior_fn=functools.partial(prior_trainable,
num_updates=num_updates),
use_bias=True,
kl_weight=1. / num_updates,
kl_use_exact=True,
name='fc1000')(x)
else:
x = layers.Dense(num_classes,
kernel_regularizer=regularizers.l2(L2_WEIGHT_DECAY),
bias_regularizer=regularizers.l2(L2_WEIGHT_DECAY),
name='fc1000')(x)
# Create model.
return models.Model(img_input, x, name='resnet50')
class_mapping['bg'] = len(class_mapping)
class_mapping = {v: k for k, v in class_mapping.items()}
print(class_mapping)
class_to_color = {
class_mapping[v]: np.random.randint(0, 255, 3)
for v in class_mapping
}
C.num_rois = int(options.num_rois)
if C.network == 'resnet50':
num_features = 1024
elif C.network == 'vgg':
num_features = 512
if K.image_data_format() == 'channels_first':
input_shape_img = (3, None, None)
input_shape_features = (num_features, None, None)
else:
input_shape_img = (None, None, 3)
input_shape_features = (None, None, num_features)
img_input = Input(shape=input_shape_img)
roi_input = Input(shape=(C.num_rois, 4))
feature_map_input = Input(shape=input_shape_features)
# define the base network (resnet here, can be VGG, Inception, etc)
shared_layers = nn.nn_base(img_input, trainable=True)
# define the RPN, built on the base layers
num_anchors = len(C.anchor_box_scales) * len(C.anchor_box_ratios)