def TinyYOLO(inputs, num_anchors, num_classes):
'''Create Tiny YOLO3 Model in Keras'''
x1 = compose(
YOLOConv2D_BN_Leaky(16, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
YOLOConv2D_BN_Leaky(32, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
YOLOConv2D_BN_Leaky(64, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
YOLOConv2D_BN_Leaky(128, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
YOLOConv2D_BN_Leaky(256, (3, 3)))(inputs)
x2 = compose(
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
YOLOConv2D_BN_Leaky(512, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(1, 1), padding='same'),
YOLOConv2D_BN_Leaky(1024, (3, 3)), YOLOConv2D_BN_Leaky(256,
(1, 1)))(x1)
y1 = compose(YOLOConv2D_BN_Leaky(512, (3, 3)),
YOLOConv2D(num_anchors * (num_classes + 5), (1, 1)))(x2)
x2 = compose(YOLOConv2D_BN_Leaky(128, (1, 1)), UpSampling2D(2))(x2)
y2 = compose(Concatenate(), YOLOConv2D_BN_Leaky(256, (3, 3)),
YOLOConv2D(num_anchors * (num_classes + 5), (1, 1)))([x2, x1])
return Model(inputs, [y1, y2])
def tiny_yolo_body(inputs, num_anchors, num_classes):
'''Create Tiny YOLO_v3 model CNN body in keras.'''
x1 = compose(
DarknetConv2D_BN_Leaky(16, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
DarknetConv2D_BN_Leaky(32, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
DarknetConv2D_BN_Leaky(64, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
DarknetConv2D_BN_Leaky(128, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
DarknetConv2D_BN_Leaky(256, (3, 3)))(inputs)
x2 = compose(
MaxPooling2D(pool_size=(2, 2), strides=(2, 2), padding='same'),
DarknetConv2D_BN_Leaky(512, (3, 3)),
MaxPooling2D(pool_size=(2, 2), strides=(1, 1), padding='same'),
DarknetConv2D_BN_Leaky(1024, (3, 3)),
DarknetConv2D_BN_Leaky(256, (1, 1)))(x1)
y1 = compose(DarknetConv2D_BN_Leaky(512, (3, 3)),
DarknetConv2D(num_anchors * (num_classes + 5), (1, 1)))(x2)
x2 = compose(DarknetConv2D_BN_Leaky(128, (1, 1)), UpSampling2D(2))(x2)
y2 = compose(Concatenate(), DarknetConv2D_BN_Leaky(256, (3, 3)),
DarknetConv2D(num_anchors * (num_classes + 5),
(1, 1)))([x2, x1])
return Model(inputs, [y1, y2])
def make_last_layers(x, num_filters, out_filters):
'''6 Conv2D_BN_Leaky layers followed by a Conv2D_linear layer'''
x = compose(DarknetConv2D_BN_Leaky(num_filters, (1, 1)),
DarknetConv2D_BN_Leaky(num_filters * 2, (3, 3)),
DarknetConv2D_BN_Leaky(num_filters, (1, 1)),
DarknetConv2D_BN_Leaky(num_filters * 2, (3, 3)),
DarknetConv2D_BN_Leaky(num_filters, (1, 1)))(x)
y = compose(DarknetConv2D_BN_Leaky(num_filters * 2, (3, 3)),
DarknetConv2D(out_filters, (1, 1)))(x)
return x, y
def UpperLayer(x, num_filters, out_filters):
'''6 Conv2D_BN_Leaky layers followed by Conv2D_Linear layer'''
x = compose(YOLOConv2D_BN_Leaky(num_filters, (1, 1)),
YOLOConv2D_BN_Leaky(num_filters * 2, (3, 3)),
YOLOConv2D_BN_Leaky(num_filters, (1, 1)),
YOLOConv2D_BN_Leaky(num_filters * 2, (3, 3)),
YOLOConv2D_BN_Leaky(num_filters, (1, 1)))(x)
y = compose(YOLOConv2D_BN_Leaky(num_filters * 2, (3, 3)),
YOLOConv2D(out_filters, (1, 1)))(x)
return x, y
def yolo_body(inputs, num_anchors, num_classes):
"""Create YOLO_V3 model CNN body in Keras."""
darknet = Model(inputs, darknet_body(inputs))
x, y1 = make_last_layers(darknet.output, 512,
num_anchors * (num_classes + 5))
x = compose(DarknetConv2D_BN_Leaky(256, (1, 1)), UpSampling2D(2))(x)
x = Concatenate()([x, darknet.layers[152].output])
x, y2 = make_last_layers(x, 256, num_anchors * (num_classes + 5))
x = compose(DarknetConv2D_BN_Leaky(128, (1, 1)), UpSampling2D(2))(x)
x = Concatenate()([x, darknet.layers[92].output])
x, y3 = make_last_layers(x, 128, num_anchors * (num_classes + 5))
return Model(inputs, [y1, y2, y3])
def DarknetConv2D_BN_Leaky(*args, **kwargs):
"""Darknet Convolution2D followed by BatchNormalization and LeakyReLU."""
no_bias_kwargs = {'use_bias': False}
no_bias_kwargs.update(kwargs)
return compose(
DarknetConv2D(*args, **no_bias_kwargs),
BatchNormalization(),
LeakyReLU(alpha=0.1))
def YOLOBody(inputs, num_anchors, num_classes):
'''Create YOLO3 Model in Keras'''
model = Model(inputs, Body(inputs))
x, y1 = UpperLayer(model.output, 512, num_anchors * (num_classes + 5))
x = compose(YOLOConv2D_BN_Leaky(256, (1, 1)), UpSampling2D(2))(x)
x = Concatenate()([x, model.layers[152].output])
x, y2 = UpperLayer(x, 256, num_anchors * (num_classes + 5))
x = compose(YOLOConv2D_BN_Leaky(128, (1, 1)), UpSampling2D(2))(x)
x = Concatenate()([x, model.layers[92].output])
x, y3 = UpperLayer(x, 128, num_anchors * (num_classes + 5))
return Model(inputs, [y1, y2, y3])
def resblock_body(x, num_filters, num_blocks):
'''A series of resblocks starting with a downsampling Convolution2D'''
# Darknet uses left and top padding instead of 'same' mode
x = ZeroPadding2D(((1, 0), (1, 0)))(x)
x = DarknetConv2D_BN_Leaky(num_filters, (3, 3), strides=(2, 2))(x)
for i in range(num_blocks):
y = compose(DarknetConv2D_BN_Leaky(num_filters // 2, (1, 1)),
DarknetConv2D_BN_Leaky(num_filters, (3, 3)))(x)
x = Add()([x, y])
return x
def ResBlock(x, num_filters, num_blocks):
x = ZeroPadding2D(((1, 0), (1, 0)))(x)
x = YOLOConv2D_BN_Leaky(num_filters, (3, 3), strides=(2, 2))(x)
for i in range(num_blocks):
y = compose(YOLOConv2D_BN_Leaky(num_filters // 2, (1, 1)),
YOLOConv2D_BN_Leaky(num_filters, (3, 3)))(x)
x = Add()([x, y])
return x
def yolo_body_distributed(inputs, num_anchors, num_classes):
"""Create YOLO_V3 model CNN body in Keras."""
darknet = Model(inputs, darknet_body_distributed(inputs))
with tf.device('/gpu:1'):
x, y1 = make_last_layers_distributed(darknet.output, 512, num_anchors*(num_classes+5))
with tf.device('/gpu:0'):
x = compose(
DarknetConv2D_BN_Leaky(256, (1,1)),
UpSampling2D(2))(x)
with tf.device('/cpu:0'):
x = Concatenate()([x,darknet.layers[152].output])
with tf.device('/gpu:1'):
x, y2 = make_last_layers_distributed(x, 256, num_anchors*(num_classes+5))
with tf.device('/gpu:0'):
x = compose(
DarknetConv2D_BN_Leaky(128, (1,1)),
UpSampling2D(2))(x)
with tf.device('/cpu:0'):
x = Concatenate()([x,darknet.layers[92].output])
with tf.device('/gpu:0'):
x, y3 = make_last_layers_distributed(x, 128, num_anchors*(num_classes+5))
return Model(inputs, [y1,y2,y3])
def YOLOConv2D_BN_Leaky(*args, **kwargs):
no_bias_kwargs = {'use_bias': False}
no_bias_kwargs.update(kwargs)
return compose(YOLOConv2D(*args, **no_bias_kwargs), BatchNormalization(),
LeakyReLU(alpha=0.1))
