简单堆叠网络容易造成梯度消失或梯度爆炸，解决的方法主要有数据标准化，权重初始化和BN(Batch Normalization)层
resnet主要创新点：
①. 能够拥有超深的网络结构
②. 提出残差(residual)模块
③. 使用BN(Batch Normalization)层加速训练
1、残差结构
主分支与shotcut的输出特征矩阵shape必须一致。
残差结构图
左图应用于resnet-34, 右图残差结构应用于resnet-50/101/152
右图使用1×1的卷积用于降维和升维
不同网络的参数比较
虚线残差结构
注意 原论文中右图残差结构的主分支上，第一个1×1的卷积层步长为2，第二个3×3的卷积层是1
pytorch官方实现的代码中第一个1×1的卷积层步长为1，第二个3×3的卷积层是2，在imagenet数据集上准确率能够提升0.5%。
该结构用于改变特征图的宽和高
2、resnet-18、resnet-34使用的残差结构
BasicBlock
代码部分
class BasicBlock(nn.Module):
"""resnet-18、resnet-34所用的残差块"""
# 残差结构主分支中所采用的卷积核个数是否改变
expansion = 1
def __init__(self, in_channel, out_channel, stride = 1, downsample = None):
"""
:param in_channel: 输入特征矩阵深度
:param out_channel: 输出特征矩阵深度
:param stride:
:param downsample: 下采样
"""
super(BasicBlock, self).__init__()
self.conv1 = nn.Conv2d(in_channels=in_channel, out_channels=out_channel,
kernel_size=3, stride=stride, padding=1, bias=False)
self.bn1 = nn.BatchNorm2d(out_channel)
self.relu = nn.ReLU()
self.conv2 = nn.Conv2d(in_channels=out_channel, out_channels=out_channel,
kernel_size=3, stride=1, padding=1, bias=1)
self.bn2 = nn.BatchNorm2d(out_channel)
self.downsample = downsample
def forward(self, x):
identity = x # 捷径分支输出值
if self.downsample is not None:
identity = self.downsample(x)
# 主分支输出
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
out += identity
out = self.relu(out)
return out
3、resnet-50、resnet-101、resnet-152使用的残差结构
Bottleneck
代码部分
class Bottleneck(nn.Module):
"""resnet-50、resnet-101、resnet-152所用残差结构"""
expansion = 4
def __init__(self, in_channel, out_channel, stride = 1, downsample = None):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2d(in_channels=in_channel,out_channels=out_channel,
kernel_size=1, stride=1, bias=False)
self.bn1 = nn.BatchNorm2d(out_channel)
self.conv2 = nn.Conv2d(in_channels=out_channel, out_channels=out_channel,
kernel_size=3, stride=stride, padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(out_channel)
self.conv3 = nn.Conv2d(in_channels=out_channel, out_channels=out_channel * self.expansion,
kernel_size=1, stride=1,bias=False)
self.bn3 = nn.BatchNorm2d(out_channel * self.expansion)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
def forward(self, x):
identity = x
if self.downsample is not None:
identity = self.downsample(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)
out += identity
out = self.relu(out)
return out
4、resnet系列网络
代码部分
class Resnet(nn.Module):
def __init__(self, block, blocks_num):
"""
:param block:
:param blocks_num: 残差结构数目
"""
super(Resnet, self).__init__()
self.in_channel = 64
self.conv1 = nn.Conv2d(3, self.in_channel, kernel_size=(7, 7), padding=(3, 3), bias=False)
self.bn1 = nn.BatchNorm2d(self.in_channel)
self.relu = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, blocks_num[0])
self.layer2 = self._make_layer(block, 128, blocks_num[1], stride=2)
self.layer3 = self._make_layer(block, 256, blocks_num[2], stride=2)
self.layer4 = self._make_layer(block, 512, blocks_num[3], stride=2)
# 网络参数初始化
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode="fan_out",nonlinearity='relu')
def _make_layer(self, block, channel, block_num, stride=1):
downsample = None
if stride != 1 or self.in_channel != channel * block.expansion:
downsample = nn.Sequential(
nn.Conv2d(self.in_channel, channel * block.expansion, kernel_size=1, stride=stride,bias=False),
nn.BatchNorm2d(channel*block.expansion)
)
layers = []
layers.append(block(self.in_channel, channel, downsample=downsample, stride=stride))
self.in_channel = channel * block.expansion
for _ in range(1, block_num):
layers.append(block(self.in_channel, channel))
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)
return x
def resnet18(pretrained=True):
"""Constructs a ResNet-18 model.
Args:
pretrained (bool): If True, returns a model pre-trained on ImageNet
"""
model = Resnet(BasicBlock, [2, 2, 2, 2])
if pretrained:
model.load_state_dict(model_zoo.load_url(
model_urls['resnet18']), strict=False)
return model
if __name__ == '__main__':
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
resnet18 = resnet18(True).to(device)
print(summary(resnet18, input_size = (3, 224, 224)))
