def bn_inceptionv2_cifar_model(input, labelDim, bnTimeConst):
# 32 x 32 x 3
conv1a = conv_bn_relu_layer(input, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 32
conv1b = conv_bn_relu_layer(conv1a, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 32
conv1c = conv_bn_relu_layer(conv1b, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 64
conv2a = conv_bn_relu_layer(conv1c, 64, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 64
conv2b = conv_bn_relu_layer(conv2a, 64, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 64
conv2c = conv_bn_relu_layer(conv2b, 64, (3,3), (1,1), True, bnTimeConst)
# Inception Blocks
# 32 x 32 x 64
inception3a = inception_block1(conv2b, 32, 32, 32, 32, 48, 16, bnTimeConst)
# 32 x 32 x 128
inception3b = inception_block1(inception3a, 0, 64, 80, 32, 48, 0, bnTimeConst)
# 32 x 32 x 128
inception3c = inception_block1(inception3b, 0, 64, 80, 32, 48, 0, bnTimeConst)
# 32 x 32 x 128
pool1 = AveragePooling(filter_shape=(3,3), strides = (2,2), pad = True)(inception3c)
# 16 x 16 x 256
inception4a = inception_block2(inception3c, 96, 48, 64, 48, 64, 64, bnTimeConst)
# 16 x 16 x 256
inception4b = inception_block2(inception4a, 96, 48, 64, 48, 64, 64, bnTimeConst)
# 16 x 16 x 256
inception4c = inception_block2(inception4b, 96, 48, 64, 48, 64, 64, bnTimeConst)
# 16 x 16 x 288
inception4d = inception_block2(inception4c, 48, 64, 96, 80, 96, 64, bnTimeConst)
# 16 x 16 x 288
inception4e = inception_block2(inception4d, 0, 128, 192, 192, 256, 0, bnTimeConst)
# 16 x 16 x 288
pool2 = AveragePooling(filter_shape=(3,3), strides = (2,2), pad = True)(inception4e)
# Inception Blocks
# 8 x 8 x 512
inception5a = inception_block1(pool2, 176, 96, 160, 96, 112, 64, bnTimeConst)
# 8 x 8 x 512
inception5b = inception_block1(inception5a, 176, 96, 160, 96, 112, 64, bnTimeConst)
# Global Average
# 8 x 8 x 512
pool3 = AveragePooling(filter_shape=(8,8))(inception5a)
# 1 x 1 x 512
z = Dense(labelDim, init=he_normal())(pool3)
return z
def inception_block_3(input, num1x1, num7x7, num7x7dbl, numPool, bnTimeConst):
# 1x1 Convolution
branch1x1 = conv_bn_relu_layer(input, num1x1, (1, 1), (1, 1), True,
bnTimeConst)
# 1x1 -> 1x3 -> 3x1
branch7x7_1 = conv_bn_relu_layer(input, num7x7[0], (1, 1), (1, 1), True,
bnTimeConst)
branch7x7_2 = conv_bn_relu_layer(branch7x7_1, num7x7[1], (1, 3), (1, 1),
True, bnTimeConst)
branch7x7 = conv_bn_relu_layer(branch7x7_2, num7x7[2], (3, 1), (1, 1),
True, bnTimeConst)
# 1x1 -> 1x3 -> 3x1 -> 1x3 -> 3x1
branch7x7dbl_1 = conv_bn_relu_layer(input, num7x7dbl[0], (1, 1), (1, 1),
True, bnTimeConst)
branch7x7dbl_2 = conv_bn_relu_layer(branch7x7dbl_1, num7x7dbl[1], (3, 1),
(1, 1), True, bnTimeConst)
branch7x7dbl_3 = conv_bn_relu_layer(branch7x7dbl_2, num7x7dbl[2], (1, 3),
(1, 1), True, bnTimeConst)
branch7x7dbl_4 = conv_bn_relu_layer(branch7x7dbl_3, num7x7dbl[3], (3, 1),
(1, 1), True, bnTimeConst)
branch7x7dbl = conv_bn_relu_layer(branch7x7dbl_4, num7x7dbl[4], (1, 3),
(1, 1), True, bnTimeConst)
# Average Pooling
branchPool_avgpool = AveragePooling((3, 3), strides=(1, 1),
pad=True)(input)
branchPool = conv_bn_relu_layer(branchPool_avgpool, numPool, (1, 1),
(1, 1), True, bnTimeConst)
out = splice(branch1x1, branch7x7, branch7x7dbl, branchPool, axis=0)
return out
def create_imagenet_model_bottleneck(input, num_stack_layers, num_classes,
stride1x1, stride3x3):
c_map = [64, 128, 256, 512, 1024, 2048]
# conv1 and max pooling
conv1 = conv_bn_relu(input, (7, 7), c_map[0], strides=(2, 2))
pool1 = MaxPooling((3, 3), strides=(2, 2), pad=True)(conv1)
# conv2_x
r2_1 = resnet_bottleneck_inc(pool1, c_map[2], c_map[0], (1, 1), (1, 1))
r2_2 = resnet_bottleneck_stack(r2_1, num_stack_layers[0], c_map[2],
c_map[0])
# conv3_x
r3_1 = resnet_bottleneck_inc(r2_2, c_map[3], c_map[1], stride1x1,
stride3x3)
r3_2 = resnet_bottleneck_stack(r3_1, num_stack_layers[1], c_map[3],
c_map[1])
# conv4_x
r4_1 = resnet_bottleneck_inc(r3_2, c_map[4], c_map[2], stride1x1,
stride3x3)
r4_2 = resnet_bottleneck_stack(r4_1, num_stack_layers[2], c_map[4],
c_map[2])
# conv5_x
r5_1 = resnet_bottleneck_inc(r4_2, c_map[5], c_map[3], stride1x1,
stride3x3)
r5_2 = resnet_bottleneck_stack(r5_1, num_stack_layers[3], c_map[5],
c_map[3])
# Global average pooling and output
pool = AveragePooling(filter_shape=(7, 7), name='final_avg_pooling')(r5_2)
z = Dense(num_classes, init=C.normal(0.01))(pool)
return z
def inception_block_5(input, num1x1, num3x3, num3x3_3x3, numPool, bnTimeConst):
# 1x1 Convolution
branch1x1 = conv_bn_relu_layer(input, num1x1, (1,1), (1,1), True, bnTimeConst)
# 3x3 Convolution
branch3x3_1 = conv_bn_relu_layer(input, num3x3[0], (1,1), (1,1), True, bnTimeConst)
branch3x3_2 = conv_bn_relu_layer(branch3x3_1, num3x3[1], (1,3), (1,1), True, bnTimeConst)
branch3x3_3 = conv_bn_relu_layer(branch3x3_1, num3x3[2], (3,1), (1,1), True, bnTimeConst)
branch3x3 = splice(branch3x3_2, branch3x3_3, axis=0)
# 3x3 3x3 Convolution
branch3x3_3x3_1 = conv_bn_relu_layer(input, num3x3_3x3[0], (1,1), (1,1), True, bnTimeConst)
branch3x3_3x3_2 = conv_bn_relu_layer(branch3x3_3x3_1, num3x3_3x3[1], (3,3), (1,1), True, bnTimeConst)
branch3x3_3x3_3 = conv_bn_relu_layer(branch3x3_3x3_2, num3x3_3x3[1], (1,3), (1,1), True, bnTimeConst)
branch3x3_3x3_4 = conv_bn_relu_layer(branch3x3_3x3_2, num3x3_3x3[3], (3,1), (1,1), True, bnTimeConst)
branch3x3_3x3 = splice(branch3x3_3x3_3, branch3x3_3x3_4, axis=0)
# Average Pooling
branchPool_avgpool = AveragePooling((3,3), strides=(1,1), pad=True)(input)
branchPool = conv_bn_relu_layer(branchPool_avgpool, numPool, (1,1), (1,1), True, bnTimeConst)
out = splice(branch1x1, branch3x3, branch3x3_3x3, branchPool, axis=0)
return out
def bn_inception_cifar_model(input, labelDim, bnTimeConst):
# 32 x 32 x 3
conv1a = conv_bn_relu_layer(input, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 32
conv1b = conv_bn_relu_layer(conv1a, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 32
conv2a = conv_bn_relu_layer(conv1b, 64, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 64
conv2b = conv_bn_relu_layer(conv2a, 128, (3,3), (1,1), True, bnTimeConst)
# Inception Blocks
# 32 x 32 x 128
inception3a = inception_block_with_maxpool(conv2b, 32, 32, 32, 32, 48, 16, bnTimeConst)
# 32 x 32 x 128
inception3b = inception_block_with_maxpool(inception3a, 32, 32, 32, 32, 48, 16, bnTimeConst)
# 16 x 16 x 128
inception4a = inception_block_with_avgpool(inception3b, 96, 48, 64, 48, 64, 64, bnTimeConst)
# 16 x 16 x 288
inception4b = inception_block_with_avgpool(inception4a, 48, 64, 96, 80, 96, 64, bnTimeConst)
# 16 x 16 x 288
inception4c = inception_block_with_avgpool(inception4b, 48, 64, 96, 80, 96, 64, bnTimeConst)
# 16 x 16 x 288
inception4d = inception_block_pass_through(inception4c, 0, 128, 192, 192, 256, 0, bnTimeConst)
# 8 x 8 x 512
inception5a = inception_block_with_maxpool(inception4d, 176, 96, 160, 96, 112, 64, bnTimeConst)
# Global Average
# 8 x 8 x 512
pool1 = AveragePooling(filter_shape=(8,8))(inception5a)
# 1 x 1 x 512
z = Dense(labelDim, init=he_normal())(pool1)
return z
def inception_block_1(input, num1x1, num5x5, num3x3dbl, numPool, bnTimeConst):
# 1x1
branch1x1 = conv_bn_relu_layer(input, num1x1, (1, 1), (1, 1), True,
bnTimeConst)
# 1x1 -> 5x5
branch5x5_1 = conv_bn_relu_layer(input, num5x5[0], (1, 1), (1, 1), True,
bnTimeConst)
branch5x5 = conv_bn_relu_layer(branch5x5_1, num5x5[1], (5, 5), (1, 1),
True, bnTimeConst)
# 1x1 -> 3x3 -> 3x3
branch3x3dbl_1 = conv_bn_relu_layer(input, num3x3dbl[0], (1, 1), (1, 1),
True, bnTimeConst)
branch3x3dbl_2 = conv_bn_relu_layer(branch3x3dbl_1, num3x3dbl[1], (3, 3),
(1, 1), True, bnTimeConst)
branch3x3dbl = conv_bn_relu_layer(branch3x3dbl_2, num3x3dbl[2], (3, 3),
(1, 1), True, bnTimeConst)
# Average Pooling
branchPool_avgpool = AveragePooling((3, 3), strides=(1, 1),
pad=True)(input)
branchPool = conv_bn_relu_layer(branchPool_avgpool, numPool, (1, 1),
(1, 1), True, bnTimeConst)
out = splice(branch1x1, branch5x5, branch3x3dbl, branchPool, axis=0)
return out
def inception_block_with_avgpool(input, num1x1, num3x3r, num3x3, num3x3dblr,
num3x3dbl, numPool, bnTimeConst):
# 1x1 Convolution
branch1x1 = conv_bn_relu_layer(input, num1x1, (1, 1), (1, 1), True,
bnTimeConst)
# 3x3
branch3x3_reduce = conv_bn_relu_layer(input, num3x3r, (1, 1), (1, 1), True,
bnTimeConst)
branch3x3 = conv_bn_relu_layer(branch3x3_reduce, num3x3, (3, 3), (1, 1),
True, bnTimeConst)
# 1x1 -> 3x3 -> 3x3
branch3x3dbl_reduce = conv_bn_relu_layer(input, num3x3dblr, (1, 1), (1, 1),
True, bnTimeConst)
branch3x3dbl_conv = conv_bn_relu_layer(branch3x3dbl_reduce, num3x3dbl,
(3, 3), (1, 1), True, bnTimeConst)
branch3x3dbl = conv_bn_relu_layer(branch3x3dbl_conv, num3x3dbl, (3, 3),
(1, 1), True, bnTimeConst)
# avg pooling -> 1x1
branchPool_avgpool = AveragePooling((3, 3), strides=(1, 1),
pad=True)(input)
branchPool = conv_bn_relu_layer(branchPool_avgpool, numPool, (1, 1),
(1, 1), True, bnTimeConst)
out = splice(branch1x1, branch3x3, branch3x3dbl, branchPool, axis=0)
return out
def bn_inception_model(input, labelDim, bnTimeConst):
# 224 x 224 x 3
conv1 = conv_bn_relu_layer(input, 64, (7, 7), (2, 2), True, bnTimeConst)
# 112 x 112 x 64
pool1 = MaxPooling(filter_shape=(3, 3), strides=(2, 2), pad=True)(conv1)
# 56 x 56 x 64
conv2a = conv_bn_relu_layer(pool1, 64, (1, 1), (1, 1), True, bnTimeConst)
# 56 x 56 x 64
conv2b = conv_bn_relu_layer(conv2a, 192, (3, 3), (1, 1), True, bnTimeConst)
# 56 x 56 x 192
pool2 = MaxPooling(filter_shape=(3, 3), strides=(2, 2), pad=True)(conv2b)
# Inception Blocks
# 28 x 28 x 192
inception3a = inception_block_with_avgpool(pool2, 64, 64, 64, 64, 96, 32,
bnTimeConst)
# 28 x 28 x 256
inception3b = inception_block_with_avgpool(inception3a, 64, 64, 96, 64, 96,
64, bnTimeConst)
# 28 x 28 x 320
inception3c = inception_block_pass_through(inception3b, 0, 128, 160, 64,
96, 0, bnTimeConst)
# 14 x 14 x 576
inception4a = inception_block_with_avgpool(inception3c, 224, 64, 96, 96,
128, 128, bnTimeConst)
# 14 x 14 x 576
inception4b = inception_block_with_avgpool(inception4a, 192, 96, 128, 96,
128, 128, bnTimeConst)
# 14 x 14 x 576
inception4c = inception_block_with_avgpool(inception4b, 160, 128, 160, 128,
160, 128, bnTimeConst)
# 14 x 14 x 576
inception4d = inception_block_with_avgpool(inception4c, 96, 128, 192, 160,
192, 128, bnTimeConst)
# 14 x 14 x 576
inception4e = inception_block_pass_through(inception4d, 0, 128, 192, 192,
256, 0, bnTimeConst)
# 7 x 7 x 1024
inception5a = inception_block_with_avgpool(inception4e, 352, 192, 320, 160,
224, 128, bnTimeConst)
# 7 x 7 x 1024
inception5b = inception_block_with_maxpool(inception5a, 352, 192, 320, 192,
224, 128, bnTimeConst)
# Global Average
# 7 x 7 x 1024
pool3 = AveragePooling(filter_shape=(7, 7))(inception5b)
# 1 x 1 x 1024
z = Dense(labelDim, init=he_normal())(pool3)
return z
def inceptionv1_cifar_model2(input, labelDim, bnTimeConst):
# 32 x 32 x 3
conv1 = conv_bn_relu_layer(input, 32, (3, 3), (1, 1), True, bnTimeConst)
# 32 x 32 x 32
conv2 = conv_bn_relu_layer(conv1, 32, (3, 3), (1, 1), True, bnTimeConst)
# Inception Blocks
# 32 x 32 x 64
inception3a = inception_block_with_maxpool(conv2, 32, 32, 32, 32, 48, 16,
bnTimeConst)
# 32 x 32 x 128
inception3b = inception_block_with_maxpool(inception3a, 32, 32, 32, 32, 48,
16, bnTimeConst)
maxpool1 = MaxPooling((3, 3), strides=(2, 2), pad=True)(inception3b)
# 16 x 16 x 128
inception4a = inception_block_with_maxpool(maxpool1, 96, 48, 64, 48, 64,
64, bnTimeConst)
# 16 x 16 x 288
inception4b = inception_block_with_maxpool(inception4a, 96, 48, 64, 48, 64,
64, bnTimeConst)
# 16 x 16 x 288
inception4c = inception_block_with_maxpool(inception4b, 96, 48, 64, 48, 64,
64, bnTimeConst)
# 16 x 16 x 288
inception4d = inception_block_with_maxpool(inception4c, 96, 48, 64, 48, 64,
64, bnTimeConst)
# 16 x 16 x 288
inception4e = inception_block_with_maxpool(inception4d, 96, 48, 64, 48, 64,
64, bnTimeConst)
maxpool2 = MaxPooling((3, 3), strides=(2, 2), pad=True)(inception4e)
# 8 x 8 x 288
inception5a = inception_block_with_maxpool(inception4e, 176, 96, 160, 96,
112, 64, bnTimeConst)
# 8 x 8 x 512
inception5b = inception_block_with_maxpool(inception5a, 176, 96, 160, 96,
112, 64, bnTimeConst)
# Global Average
# 8 x 8 x 512
pool1 = AveragePooling(filter_shape=(8, 8))(inception5b)
# 1 x 1 x 512
z = Dense(labelDim, init=he_normal())(pool1)
return z
def Inception_A(input, a1, b1, c1, c2, d1, d2, bnTimeConst):
A1 = AveragePooling((3, 3), strides=(1, 1), pad=True)(input)
A2 = conv_bn_relu_layer(A1, a1, (3, 3), (1, 1), True, bnTimeConst)
B1 = conv_bn_relu_layer(input, b1, (1, 1), (1, 1), True, bnTimeConst)
C1 = conv_bn_relu_layer(input, c1, (1, 1), (1, 1), True, bnTimeConst)
C2 = conv_bn_relu_layer(C1, c2, (3, 3), (1, 1), True, bnTimeConst)
D1 = conv_bn_relu_layer(input, d1, (1, 1), (1, 1), True, bnTimeConst)
D2 = conv_bn_relu_layer(D1, d2, (3, 3), (1, 1), True, bnTimeConst)
D3 = conv_bn_relu_layer(D2, d2, (3, 3), (1, 1), True, bnTimeConst)
out = splice(A2, B1, C2, D3, axis=0)
return out
def create_resnet_model(input, num_classes):
conv = convolution_bn(input, (3, 3), 16)
r1_1 = resnet_basic_stack(conv, 16, 3)
r2_1 = resnet_basic_inc(r1_1, 32)
r2_2 = resnet_basic_stack(r2_1, 32, 2)
r3_1 = resnet_basic_inc(r2_2, 64)
r3_2 = resnet_basic_stack(r3_1, 64, 2)
# Global average pooling
pool = AveragePooling(filter_shape=(8, 8), strides=(1, 1))(r3_2)
net = Dense(num_classes, init=he_normal(), activation=None)(pool)
return net
def create_cifar10_model(input, num_stack_layers, num_classes):
c_map = [16, 32, 64]
conv = conv_bn_relu(input, (3, 3), c_map[0])
r1 = resnet_basic_stack(conv, num_stack_layers, c_map[0])
r2_1 = resnet_basic_inc(r1, c_map[1])
r2_2 = resnet_basic_stack(r2_1, num_stack_layers - 1, c_map[1])
r3_1 = resnet_basic_inc(r2_2, c_map[2])
r3_2 = resnet_basic_stack(r3_1, num_stack_layers - 1, c_map[2])
# Global average pooling and output
pool = AveragePooling(filter_shape=(8, 8), name='final_avg_pooling')(r3_2)
z = Dense(num_classes)(pool)
return z
def create_cifar10_model(input, num_classes):
c_map = [32, 64, 128]
conv = conv_bn_relu(input, (3, 3), c_map[0])
r1 = mobilenet_basic_stack(conv, 3, c_map[0])
r2_1 = mobilenet_basic_inc(r1, c_map[1])
r2_2 = mobilenet_basic_stack(r2_1, 3, c_map[1])
r3_1 = mobilenet_basic_inc(r2_2, c_map[2])
r3_2 = mobilenet_basic_stack(r3_1, 3, c_map[2])
# Global average pooling and output
pool = AveragePooling(filter_shape=(8, 8))(r3_2)
z = Dense(num_classes, init=he_normal())(pool)
return z
def create_resnet34(input, num_classes):
c_map = [64, 128, 256, 512]
num_layers = [3, 3, 5, 2]
conv = conv_bn_relu(input, (7, 7), c_map[0], (2, 2))
maxPool = MaxPooling((3, 3), (2, 2), pad=True)(conv)
r1 = resnet_basic_stack(maxPool, num_layers[0], c_map[0])
r2_1 = resnet_basic_inc(r1, c_map[1])
r2_2 = resnet_basic_stack(r2_1, num_layers[1], c_map[1])
r3_1 = resnet_basic_inc(r2_2, c_map[2])
r3_2 = resnet_basic_stack(r3_1, num_layers[2], c_map[2])
r4_1 = resnet_basic_inc(r3_2, c_map[3])
r4_2 = resnet_basic_stack(r4_1, num_layers[3], c_map[3])
# Global average pooling and output
pool = AveragePooling(filter_shape=(7, 7))(r4_2)
z = Dense(num_classes)(pool)
return z
def InceptionV4_ResNet_model(input, num_classes):
l = pre_block(input)
# 32 x 32 x 128
A1 = Res_A_stack(l,5, 64, 128)
RA = reduction_A(A1, 32, 32, 64, 64)
# 16 x 16 x 224
B1 = Res_B_stack(RA, 10, 128, 224)
# 16 x 16 x 448
RB = reduction_B(B1, 32, 64, 32, 64, 128)
# 8 x 8 x 480
C1 = Res_C_stack(RB, 5, 192, 480)
# Global average pooling and output
pool = AveragePooling(filter_shape=(8, 8), name='final_avg_pooling')(C1)
z = Dense(num_classes, init=normal(0.01))(pool)
return z
def Inception_ResNet_v2_model(input, num_classes):
l = pre_block(input)
# 32 x 32 x 128
A1 = Res_A_stack(l, 5, 32, 32, 32, 48, 64, 128)
# 32 x 32 x 128
RA = reduction_A(A1, 64, 32, 64, 64)
# 16 x 16 x 256
B1 = Res_B_stack(RA, 10, 128, 64, 96, 128, 256)
# 16 x 16 x 256
RB = reduction_B(B1, 32, 64, 32, 64, 32, 64, 128)
# 8 x 8 x 512
C1 = Res_C_stack(RB, 5, 128, 64, 128, 256, 512)
# Global average pooling and output
pool = AveragePooling(filter_shape=(8, 8))(C1)
z = Dense(num_classes, init=normal(0.01))(pool)
return z
def inception_v4_model(input, labelDim, bnTimeConst):
l = pre_block(input, bnTimeConst)
# 32 x 32
A1 = Inception_A(l, 32, 32, 32, 64, 32, 64, bnTimeConst)
A2 = Inception_A(A1, 32, 32, 32, 64, 32, 64, bnTimeConst)
A3 = Inception_A(A2, 32, 32, 32, 64, 32, 64, bnTimeConst)
A4 = Inception_A(A3, 32, 32, 32, 64, 32, 64, bnTimeConst)
# 32 x 32 x 192
RA = reduction_A(A4, 32, 32, 64, 64, bnTimeConst)
# 16 x 16 x 288
B1 = Inception_B(RA, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
B2 = Inception_B(B1, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
B3 = Inception_B(B2, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
B4 = Inception_B(B3, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
B5 = Inception_B(B4, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
B6 = Inception_B(B5, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
B7 = Inception_B(B6, 128, 32, 32, 64, 96, 32, 64, 96, bnTimeConst)
# 16 x 16 x 352
RB = reduction_B(B7, 64, 64, 96, bnTimeConst)
# 8 x 8 x 512
C1 = Inception_C(RB, 128, 128, 96, 64, 64, 128, 160, 64, bnTimeConst)
C2 = Inception_C(C1, 128, 128, 96, 64, 64, 128, 160, 64, bnTimeConst)
C3 = Inception_C(C2, 128, 128, 96, 64, 64, 128, 160, 64, bnTimeConst)
# 8 x 8 x 512
pool1 = AveragePooling(filter_shape=(8, 8))(C3)
# 1 x 1 x 512
z = Dense(labelDim, init=he_normal())(pool1)
return z
def inception_v3_cifar_model(input, labelDim, bnTimeConst):
# 32 x 32 x 3
conv1 = conv_bn_relu_layer(input, 32, (3, 3), (1, 1), True, bnTimeConst)
# 32 x 32 x 32
conv2 = conv_bn_relu_layer(conv1, 32, (3, 3), (1, 1), True, bnTimeConst)
# 32 x 32 x 32
conv3 = conv_bn_relu_layer(conv2, 64, (3, 3), (1, 1), True, bnTimeConst)
# 32 x 32 x 64
conv4 = conv_bn_relu_layer(conv3, 80, (1, 1), (1, 1), True, bnTimeConst)
# 32 x 32 x 80
conv5 = conv_bn_relu_layer(conv4, 128, (3, 3), (1, 1), True, bnTimeConst)
# 32 x 32 x 128
pool1 = MaxPooling(filter_shape=(3, 3), strides=(2, 2), pad=True)(conv5)
#
# Inception Blocks
# 16 x 16 x 128
mixed1 = inception_block_1(pool1, 32, [32, 48], [48, 64, 64], 32,
bnTimeConst)
# 16 x 16 x 160
mixed2 = inception_block_1(mixed1, 32, [32, 48], [48, 64, 64], 64,
bnTimeConst)
# 16 x 16 x 160
mixed3 = inception_block_1(mixed2, 32, [32, 48], [48, 64, 64], 64,
bnTimeConst)
# 16 x 16 x 160
#mixed4 = inception_block_2(mixed3, 32, [48, 64, 64], bnTimeConst)
mixed4 = inception_block_pass_through(mixed3, 0, 64, 80, 32, 48, 0,
bnTimeConst)
# 8 x 8 x 256
mixed5 = inception_block_3(mixed4, 192, [48, 64, 64],
[128, 128, 128, 128, 192], 192, bnTimeConst)
# 8 x 8 x
mixed6 = inception_block_3(mixed5, 192, [160, 160, 192],
[160, 160, 160, 160, 192], 192, bnTimeConst)
# 8 x 8 x 768
mixed7 = inception_block_3(mixed6, 192, [160, 160, 192],
[160, 160, 160, 160, 192], 192, bnTimeConst)
# 8 x 8 x 768
mixed8 = inception_block_3(mixed7, 192, [192, 192, 192],
[192, 192, 192, 192, 192], 192, bnTimeConst)
# 8 x 8 x 768
mixed9 = inception_block_3(mixed8, 192, [192, 192, 192],
[192, 192, 192, 192, 192], 192, bnTimeConst)
# 8 x 8 x 1280
mixed10 = inception_block_5(mixed9, 320, [384, 384, 384],
[448, 384, 384, 384], 192, bnTimeConst)
# 8 x 8 x 2048
mixed11 = inception_block_5(mixed10, 320, [384, 384, 384],
[448, 384, 384, 384], 192, bnTimeConst)
# 8 x 8 x 2048
#
# Prediction
#
pool3 = AveragePooling(filter_shape=(8, 8))(mixed11)
# 1 x 1 x 2048
drop = Dropout(dropout_rate=0.2)(pool3)
# 1 x 1 x 2048
z = Dense(labelDim, init=he_normal())(drop)
#
# Auxiliary
#
# 8 x 8 x 768
auxPool = AveragePooling(filter_shape=(3, 3), strides=(1, 1),
pad=True)(mixed8)
# 5 x 5 x 768
auxConv1 = conv_bn_relu_layer(auxPool, 128, (1, 1), (1, 1), True,
bnTimeConst)
# 5 x 5 x 128
auxConv2 = conv_bn_relu_layer(auxConv1, 256, (3, 3), (1, 1), True,
bnTimeConst)
# 1 x 1 x 768
aux = Dense(labelDim, init=he_normal())(auxConv2)
return {'z': z, 'aux': aux}
def inception_v3_cifar_model(input, labelDim, bnTimeConst):
# 32 x 32 x 3
conv1 = conv_bn_relu_layer(input, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 32
conv2 = conv_bn_relu_layer(conv1, 32, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 32
conv3 = conv_bn_relu_layer(conv2, 64, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 64
conv4 = conv_bn_relu_layer(conv3, 80, (1,1), (1,1), True, bnTimeConst)
# 32 x 32 x 80
conv5 = conv_bn_relu_layer(conv4, 128, (3,3), (1,1), True, bnTimeConst)
# 32 x 32 x 128
pool1 = MaxPooling(filter_shape=(3,3), strides=(2,2), pad=True)(conv5)
#
# Inception Blocks
# 16 x 16 x 128
mixed1 = inception_block_1(pool1, 32, [32, 48], [48, 64, 64], 32, bnTimeConst)
# 16 x 16 x 176
mixed2 = inception_block_1(mixed1, 32, [32, 48], [48, 64, 64], 32, bnTimeConst)
# 16 x 16 x 176
mixed3 = inception_block_1(mixed2, 32, [32, 48], [48, 64, 64], 32, bnTimeConst)
# 16 x 16 x 176
mixed4 = inception_block_pass_through(mixed3, 0, 32, 48, 32, 48, 0, bnTimeConst)
# 8 x 8 x 256
mixed5 = inception_block_3(mixed4, 64, [48, 64, 64], [48, 48, 48, 48, 64], 64, bnTimeConst)
# 8 x 8 x 256
mixed6 = inception_block_3(mixed5, 64, [48, 64, 64], [48, 48, 48, 48, 64], 64, bnTimeConst)
# 8 x 8 x 256
mixed7 = inception_block_3(mixed6, 64, [48, 64, 64], [48, 48, 48, 48, 64], 64, bnTimeConst)
# 8 x 8 x 256
mixed8 = inception_block_3(mixed7, 80, [48, 64, 64], [48, 48, 48, 48, 64], 80, bnTimeConst)
# 8 x 8 x 288
mixed9 = inception_block_3(mixed8, 128, [64, 128, 128], [64, 64, 64, 64, 128], 128, bnTimeConst)
# 8 x 8 x 512
mixed10 = inception_block_5(mixed9, 128, [64, 128, 128], [64, 64, 64, 128], 128, bnTimeConst)
# 8 x 8 x 512
mixed11 = inception_block_5(mixed10, 128, [64, 128, 128], [64, 64, 64, 128], 128, bnTimeConst)
# 8 x 8 x 512
#
# Prediction
#
pool2 = AveragePooling(filter_shape=(8,8))(mixed11)
# 1 x 1 x 512
z = Dense(labelDim, init=he_normal())(pool2)
#
# Auxiliary
#
# 8 x 8 x 288
auxPool = AveragePooling(filter_shape=(3,3), strides=(1,1), pad=True)(mixed8)
# 8 x 8 x 288
auxConv1 = conv_bn_relu_layer(auxPool, 320, (1,1), (1,1), True, bnTimeConst)
# 8 x 8 x 320
auxConv2 = conv_bn_relu_layer(auxConv1, 512, (3,3), (1,1), True, bnTimeConst)
# 8 x 8 x 512
aux = Dense(labelDim, init=he_normal())(auxConv2)
return {
'z': z,
'aux': aux
}
def clone_cntk_layer(self, feature):
"""Returns a clone of the CNTK layer for per-layer forward prop"""
pad, filterShape, stride = self.get_cntk_parameters()
return AveragePooling(filterShape, strides=(stride, stride),
pad=pad)(feature)
def inception_v3_model(input, labelDim, dropRate, bnTimeConst):
# 299 x 299 x 3
conv1 = conv_bn_relu_layer(input, 32, (3, 3), (2, 2), False, bnTimeConst)
# 149 x 149 x 32
conv2 = conv_bn_relu_layer(conv1, 32, (3, 3), (1, 1), False, bnTimeConst)
# 147 x 147 x 32
conv3 = conv_bn_relu_layer(conv2, 64, (3, 3), (1, 1), True, bnTimeConst)
# 147 x 147 x 64
pool1 = MaxPooling(filter_shape=(3, 3), strides=(2, 2), pad=False)(conv3)
# 73 x 73 x 64
conv4 = conv_bn_relu_layer(pool1, 80, (1, 1), (1, 1), False, bnTimeConst)
# 73 x 73 x 80
conv5 = conv_bn_relu_layer(conv4, 192, (3, 3), (1, 1), False, bnTimeConst)
# 71 x 71 x 192
pool2 = MaxPooling(filter_shape=(3, 3), strides=(2, 2), pad=False)(conv5)
# 35 x 35 x 192
#
# Inception Blocks
#
mixed1 = inception_block_1(pool2, 64, [48, 64], [64, 96, 96], 32,
bnTimeConst)
# 35 x 35 x 256
mixed2 = inception_block_1(mixed1, 64, [48, 64], [64, 96, 96], 64,
bnTimeConst)
# 35 x 35 x 288
mixed3 = inception_block_1(mixed2, 64, [48, 64], [64, 96, 96], 64,
bnTimeConst)
# 35 x 35 x 288
mixed4 = inception_block_2(mixed3, 384, [64, 96, 96], bnTimeConst)
# 17 x 17 x 768
mixed5 = inception_block_3(mixed4, 192, [128, 128, 192],
[128, 128, 128, 128, 192], 192, bnTimeConst)
# 17 x 17 x 768
mixed6 = inception_block_3(mixed5, 192, [160, 160, 192],
[160, 160, 160, 160, 192], 192, bnTimeConst)
# 17 x 17 x 768
mixed7 = inception_block_3(mixed6, 192, [160, 160, 192],
[160, 160, 160, 160, 192], 192, bnTimeConst)
# 17 x 17 x 768
mixed8 = inception_block_3(mixed7, 192, [192, 192, 192],
[192, 192, 192, 192, 192], 192, bnTimeConst)
# 17 x 17 x 768
mixed9 = inception_block_4(mixed8, [192, 320], [192, 192, 192, 192],
bnTimeConst)
# 8 x 8 x 1280
mixed10 = inception_block_5(mixed9, 320, [384, 384, 384],
[448, 384, 384, 384], 192, bnTimeConst)
# 8 x 8 x 2048
mixed11 = inception_block_5(mixed10, 320, [384, 384, 384],
[448, 384, 384, 384], 192, bnTimeConst)
# 8 x 8 x 2048
#
# Prediction
#
pool3 = AveragePooling(filter_shape=(8, 8), pad=False)(mixed11)
# 1 x 1 x 2048
drop = Dropout(dropout_rate=dropRate)(pool3)
# 1 x 1 x 2048
z = Dense(labelDim, init=he_normal())(drop)
#
# Auxiliary
#
# 17 x 17 x 768
auxPool = AveragePooling(filter_shape=(5, 5), strides=(3, 3),
pad=False)(mixed8)
# 5 x 5 x 768
auxConv1 = conv_bn_relu_layer(auxPool, 128, (1, 1), (1, 1), True,
bnTimeConst)
# 5 x 5 x 128
auxConv2 = conv_bn_relu_layer(auxConv1, 768, (5, 5), (1, 1), False,
bnTimeConst)
# 1 x 1 x 768
aux = Dense(labelDim, init=he_normal())(auxConv2)
return {'z': z, 'aux': aux}
