首先前四层仍是resnet50的前四层，其中第三层的第一个block会将feature map下采样2倍，第四层的第一个block会将feature map下采样2倍。代码实现如下所示：
class Bottleneck(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(Bottleneck, self).__init__()
#1x1卷积核的作用就是改变通道数，所以不用在乎padding和stride。此时，通道数由inplanes变为planes。
self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
#3x3卷积核可能改变输出尺寸大小，输出尺寸大小=输入尺寸大小/stride。通道数仍为planes。
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(planes)
# 1x1卷积核的作用就是改变通道数，所以不用在乎padding和stride。此时，通道数为4*planes。
self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
self.bn3 = nn.BatchNorm2d(planes * 4)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None:
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
class DetNet(nn.Module):
def __init__(self, block, layers, num_classes=1000):
self.inplanes = 64
super(DetNet, self).__init__()
self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
bias=False)
self.bn1 = nn.BatchNorm2d(64)
self.relu = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
#layers:[3, 4, 6, 3, 3]
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
self.layer4 = self._make_new_layer(256, layers[3])
self.layer5 = self._make_new_layer(256, layers[4])
self.avgpool = nn.AdaptiveAvgPool2d(1)
self.fc = nn.Linear(1024, num_classes)
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, math.sqrt(2. / n))
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
def _make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
nn.Conv2d(self.inplanes, planes * block.expansion,
kernel_size=1, stride=stride, bias=False),
nn.BatchNorm2d(planes * block.expansion),
)
layers = []
#每个layers的第一个block的第二层可能改变feature map大小，具体由stride确定
#其他block的第二层不改变feature map大小
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for i in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
def _make_new_layer(self, planes, blocks):
downsample = None
block_b = BottleneckB
block_a = BottleneckA
layers = []
layers.append(block_b(self.inplanes, planes, stride=1, downsample=downsample))
self.inplanes = planes * block_b.expansion
for i in range(1, blocks):
layers.append(block_a(self.inplanes, planes))
return nn.Sequential(*layers)
def forward(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.maxpool(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.layer4(x)
x = self.layer5(x)
x = self.avgpool(x)
x = x.view(x.size(0), -1)
x = self.fc(x)
return x
和resnet50相比，此代码增加了_make_new_layer()函数，用于构造额外的layer。观察此函数，可以发现，多了两个新block，分别是BottleneckB和BottleneckA。它们和Bottleneck基本一样，只是在第二层3x3的卷积层增加了空洞卷积操作。
BottleneckA代码实现如下：
class BottleneckA(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(BottleneckA, self).__init__()
#inplanes必须是planes的4倍
assert inplanes == (planes * 4), 'inplanes != planes * 4'
#stride必须是1
assert stride == 1, 'stride != 1'
#downsample必须是空
assert downsample is None, 'downsample is not None'
#这个新的模块和resnet的基本模块非常地相似，可以说是几乎一模一样。
# 只是第二层的3x3卷积操作加了空洞卷积而已。
self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False) # inplanes = 1024, planes = 256
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1, dilation=2,
padding=2, bias=False) # stride = 1, dilation = 2
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
self.bn3 = nn.BatchNorm2d(planes * 4)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None: # downsample always is None, because stride=1 and inplanes=expansion * planes
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
BottleneckB代码实现如下：
class BottleneckB(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(BottleneckB, self).__init__()
assert inplanes == (planes * 4), 'inplanes != planes * 4'
assert stride == 1, 'stride != 1'
assert downsample is None, 'downsample is not None'
# 这个新的模块和resnet的基本模块非常地相似，可以说是几乎一模一样。只是第二层的3x3卷积操作加了空洞卷积而已。
# shortcut支路增加了1x1的卷积操作
self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False) # inplanes = 1024, planes = 256
self.bn1 = nn.BatchNorm2d(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1, dilation=2,
padding=2, bias=False) # stride = 1, dilation = 2
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
self.bn3 = nn.BatchNorm2d(planes * 4)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
self.extra_conv = nn.Sequential(
nn.Conv2d(inplanes, planes * 4, kernel_size=1, bias=False),
nn.BatchNorm2d(planes * 4)
)
def forward(self, x):
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
residual = self.extra_conv(x)
if self.downsample is not None: # downsample always is None, because stride=1 and inplanes=expansion * planes
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out