def yolo_body(inputs, num_anchors, num_classes):
feat1, feat2, feat3 = mobilenetv1(inputs)
# (13, 13, 1024) -> (13, 13, 512)
x, y1 = make_last_layers(feat3, 512, num_anchors*(num_classes + 5))
# (13, 13, 512) -> (26, 26, 256)
x = compose(
DarknetConv2D_BN_Leaky(256, (1, 1)),
UpSampling2D(2))(x)
x = Concatenate()([x, feat2])
# (26, 26, 256) -> (52, 52, 128)
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, feat1])
# (52, 52, 256) -> (52, 52, 128)
x, y3 = make_last_layers(x, 128, num_anchors*(num_classes + 5))
return Model(inputs, [y1, y2, y3])
def make_last_layers(inputs, filters):
x = DarknetConv2D_BN_Leaky(filters, (1, 1))(inputs)
x = DarknetConv2D_BN_Leaky(filters * 2, (3, 3))(x)
x = DarknetConv2D_BN_Leaky(filters, (1, 1))(x)
x = DarknetConv2D_BN_Leaky(filters * 2, (3, 3))(x)
x = DarknetConv2D_BN_Leaky(filters, (1, 1))(x)
return x
def yolo_body(inputs, num_anchors, num_classes):
feat1, feat2, feat3 = CSPdarknet53(inputs)
# SPP net
P5 = DarknetConv2D_BN_Leaky(512, (1, 1))(feat3)
P5 = DarknetConv2D_BN_Leaky(1024, (3, 3))(P5)
P5 = DarknetConv2D_BN_Leaky(512, (1, 1))(P5)
maxpool1 = MaxPooling2D((13, 13), strides=(1, 1), padding='same')(P5)
maxpool2 = MaxPooling2D((9, 9), strides=(1, 1), padding='same')(P5)
maxpool3 = MaxPooling2D((5, 5), strides=(1, 1), padding='same')(P5)
P5 = Concatenate()([maxpool1, maxpool2, maxpool3, P5])
P5 = DarknetConv2D_BN_Leaky(512, (1, 1))(P5)
P5 = DarknetConv2D_BN_Leaky(1024, (3, 3))(P5)
P5 = DarknetConv2D_BN_Leaky(512, (1, 1))(P5)
P5_upsample = compose(DarknetConv2D_BN_Leaky(256, (1, 1)),
UpSampling2D(2))(P5)
P4 = DarknetConv2D_BN_Leaky(256, (1, 1))(feat2)
P4 = Concatenate()([P4, P5_upsample])
P4 = make_last_layers(P4, 256)
P4_upsample = compose(DarknetConv2D_BN_Leaky(128, (1, 1)),
UpSampling2D(2))(P4)
P3 = DarknetConv2D_BN_Leaky(128, (1, 1))(feat1)
P3 = Concatenate()([P3, P4_upsample])
P3 = make_last_layers(P3, 128)
# yolo_output3 shape(b, 52, 52, 3, num_classes + 5)
P3_output = DarknetConv2D_BN_Leaky(256, (3, 3))(P3)
P3_output = DarknetConv2D(num_anchors * (num_classes + 5),
(1, 1))(P3_output)
P3_downsample = ZeroPadding2D(((1, 0), (1, 0)))(P3)
P3_downsample = DarknetConv2D_BN_Leaky(256, (3, 3),
strides=(2, 2))(P3_downsample)
P4 = Concatenate()([P3_downsample, P4])
P4 = make_last_layers(P4, 256)
# yolo_output2 shape(b, 26, 26, 3, num_classes + 5)
P4_output = DarknetConv2D_BN_Leaky(512, (3, 3))(P4)
P4_output = DarknetConv2D(num_anchors * (num_classes + 5),
(1, 1))(P4_output)
P4_downsample = ZeroPadding2D(((1, 0), (1, 0)))(P4)
P4_downsample = DarknetConv2D_BN_Leaky(512, (3, 3),
strides=(2, 2))(P4_downsample)
P5 = Concatenate()([P4_downsample, P5])
P5 = make_last_layers(P5, 512)
# yolo_output1 shape(b, 13, 13, 3, num_classes + 5)
P5_output = DarknetConv2D_BN_Leaky(1024, (3, 3))(P5)
P5_output = DarknetConv2D(num_anchors * (num_classes + 5),
(1, 1))(P5_output)
return Model(inputs, [P5_output, P4_output, P3_output])