一、网络结构
1.BasicBlock
2.BootleNeck和group convolution
3.Dilated Convolution with a 3 x 3 kernel and dilation rate 2
二、torchvision源码
源码地址为：torchvision.models.resnet
resnetxt50_32x4d和resnext101_32x8d见论文Aggregated Residual Transformations for Deep Neural Networks
以下我做了小小的修改.
import torch
import torch.nn as nn
try:
from torch.hub import load_state_dict_from_url
except ImportError:
from torch.utils.model_zoo import load_url as load_state_dict_from_url
__all__ = ['ResNet', 'resnet18', 'resnet34', 'resnet50', 'resnet101',
'resnet152', 'resnext50_32x4d', 'resnext101_32x8d',
'wide_resnet50_2', 'wide_resnet101_2']
model_urls = {
'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',
'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',
'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',
'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',
'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',
'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',
'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',
'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',
'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth',
}
def conv3x3(in_planes, planes, stride=1, groups=1, dilation=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
def conv1x1(in_planes, planes, stride=1, groups=1, dilation=1):
"""1x1 convolution with padding"""
return nn.Conv2d(in_planes, planes, kernel_size=1, stride=stride, bias=False)
class BasicBlock(nn.Module): #输出channel和输入chnnel都是64
expansion = 1
__constants__ = ['downsample']
def __init__(self, in_planes, planes, stride=1, downsample=None, groups=1,
base_width=64, dialtion=1, norm_layer=None):
super(BasicBlock, self).__init__()
if norm_layer is None:
norm_layer = nn.BatchNorm2d
if groups!=1 or base_width!=64:
raise ValueError('BasicBlock only supports groups=1 and basic_width=64')
if dialtion > 1:
raise NotImplementedError("Dilation > 1 not supported in BasicBlock")
self.conv1 = conv3x3(in_planes, planes, stride=stride) #若有下采样，则在第一个3x3卷积进行
self.bn1 = norm_layer(planes)
self.relu = nn.ReLU(inplace=True)
self.conv2 = conv3x3(planes, planes)
self.bn2 = norm_layer(planes)
self.downsample = downsample
self.stride = stride
def forward(self, x):
identity = x
x = self.conv1(x)
x = self.bn1(x)
x = self.ReLU1(x)
x = self.conv2(x)
x = self.bn2(x)
if self.downsample is not None:
identity = self.downsample(identity)
out = x + identity #残差结构
out = self.relu(out)
return out
class BottleNeck(nn.Module):
expansion = 4 #in_planes=256, planes=64, 恢复时需要扩展为4倍
__constants__ = ['downsample']
def __init__(self, in_planes, planes, stride=1, downsample=None,
groups=1, base_width=64, dilation=1, norm_layer=None):
super(BasicBlock, self).__init__()
if norm_layer is None:
norm_layer = nn.BatchNorm2d
width = int(planes * (base_width / 64.)) * groups #??
self.conv1 = conv1x1(in_planes, width)
self.bn1 = norm_layer(width)
self.relu = nn.ReLU(inplace=True)
self.conv2 = conv3x3(width, width, stride=stride, groups=groups, dilation=dilation)
self.bn2 = norm_layer(width)
self.conv3 = conv1x1(width, planes * self.expansion)
self.bn3 = norm_layer(planes * self.expansion)
self.downsample = downsample
self.stride = stride
def forward(self, x):
identity = x
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.conv2(x)
x = self.bn2(x)
x = self.relu(x)
x = self.conv3(x)
x = self.bn3(x)
if self.downsample is not None:
identity = self.downsample(identity)
out = identity + x
return out
class ResNet(nn.Module):
def __init__(self, block, layers, num_classes=1000, zero_init_residual=False,
groups=1, width_per_group=64, replace_stride_with_dilation=None,
norm_layer=None):
super(ResNet, self).__init__()
if norm_layer is None:
norm_layer = nn.BatchNorm2d
self._norm_layer = norm_layer
self.inplanes = 64
self.dilation = 1
if replace_stride_with_dilation is None:
#tuple中的每个元素表示 是否用空洞卷积dilated convolution代替2x2 stride
replace_stride_with_dilation = [False, False, False]
if len(replace_stride_with_dilation) != 3:
raise ValueError("replace_stride_with_dilation should be None "
"or a 3-element tuple, got {}".format(replace_stride_with_dilation))
self.group = groups
self.base_width = width_per_group
#224x224=>112x112 64@7x7, stride 2
self.conv1 = nn.Conv2d(3, self.inplanes, kernel_size=7, stride=2, padding=3, bias=False)
self.bn1 = norm_layer(self.inplanes)
self.relu = nn.ReLU(inplace=True)
#112x112=>56x56 3x3 maxpoll, stride=2
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2, dilate=replace_stride_with_dilation[0])
self.layer3 = self._make_layer(block, 256, layers[2], stride=2, dilate=replace_stride_with_dilation[1])
self.layer4 = self._make_layer(block, 512, layers[3], stride=2, dilate=replace_stride_with_dilation[2])
self.avgpool = nn.AdaptiveAvgPool2d((1,1))
self.fc = nn.Linear(512 * block.expansion, num_classes)
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
if zero_init_residual:
for m in self.modules():
if isinstance(m, BottleNeck):
nn.init.constant_(m.bn3.weight, 0)
elif isinstance(m, BasicBlock):
nn.init.constant_(m.bn2.weight, 0)
def _make_layer(self, block, planes, blocks, stride=1, dilate=False):
norm_layer = self._norm_layer
downsample = None
prevision_dilation = self.dilation
if dilate:
self.dilation *= stride
stride = 1
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
conv1x1(self.inplanes, planes * block.expansion, stride),
norm_layer(planes * block.expansion)
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample, self.groups,
self.base_width, prevision_dilation, norm_layer))
self.inplanes = planes * block.expansion
for _ in range(1, blocks):
layers.append(block(self.inplanes, planes, groups=self.groups, bash_width=self.base_width,
dilation=self.dilation, norm_layer=norm_layer))
return nn.Sequential(*layers)
def _foward_impl(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.avgpool(x)
x = torch.flatten(x, 1)
x = self.fc(x)
return x
def forward(self, x):
return self._foward_impl(x)
def _resnet(arch, block, layers, pretiraned, progress, **kwargs):
model = ResNet(block, layers, **kwargs)
if pretiraned:
state_dict = load_state_dict_from_url(model_urls[arch], progress=progress)
model.load_state_dict(state_dict)
return model
def resnet18(pretrained=False, progress=True, **kwargs):
return _resnet('resnet18', BasicBlock, [2, 2, 2, 2], pretrained, progress, **kwargs)
def resnet34(pretrained=False, progress=True, **kwargs):
return _resnet('resnet34', BasicBlock, [3, 4, 6, 3], pretrained, progress, **kwargs)
def resnet50(pretrained=False, progress=True, **kwargs):
return _resnet('resnet50', BottleNeck, [3, 4, 6, 3], pretrained, progress, **kwargs)
def resnet101(pretrained=False, progress=True, **kwargs):
return _resnet('resnet101', BottleNeck, [3, 4, 23, 3], pretrained, progress, **kwargs)
def resnet152(pretrained=False, progress=True, **kwargs):
return _resnet('resnet152', BottleNeck, [3, 8, 36, 3], pretrained, progress, **kwargs)
def resnetxt50_32x4d(pretrained=False, progress=True, **kwargs):
kwargs['groups'] = 32
kwargs['width_per_group'] = 4
return _resnet('resnext50_32x4d', BottleNeck, [3, 4, 6, 3], pretrained, progress, **kwargs)
def resnext101_32x8d(pretrained=False, progress=True, **kwargs):
kwargs['groups'] = 32
kwargs['width_per_group'] = 8
return _resnet('resnext101_32x8d', BottleNeck, [3, 4, 23, 3],
pretrained, progress, **kwargs)
def wide_resnet50_2(pretrained=False, progress=True, **kwargs):
kwargs['width_per_group'] = 64 * 2
return _resnet('wide_resnet50_2', BottleNeck, [3, 4, 6, 3],
pretrained, progress, **kwargs)
def wide_resnet101_2(pretrained=False, progress=True, **kwargs):
kwargs['width_per_group'] = 64 * 2
return _resnet('wide_resnet101_2', BottleNeck, [3, 4, 23, 3],
pretrained, progress, **kwargs)
三、一些值得学习的用法笔记
3.1 分组卷积和空洞卷积
nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
# groups:是否采用分组卷积,即将channel分组进行卷积，再拼接.
# 默认groups=1,即正常卷积
# group=k, k需要整除in_planes和输出channel，参数量节省k倍.
# group=in_planes时，即每个通道单独进行，即为深度卷积depthwise.
# dilation:是否采用空洞卷积，即卷积核之间是否存在dilation-1的空洞.
# 默认dilaton=1,即为正常卷积
3.2 先搭block,再搭网络的方式
class BasicBlock(nn.Module):
...
class BottleNeck(nn.Module):
...
#先定义block,再堆叠block的方法，值得我们学习
3.3 抛出错误
if groups!=1 or base_width!=64:
raise ValueError('BasicBlock only supports groups=1 and basic_width=64')
#当出现可以预想的错误时，采用抛出错误的方法，即rasie Error('erro type')提倡使用
3.4 残差结构写法
#残差结构的写法
def forward(self, x):
identity = x
x = self.conv(x)
if self.downsample is not None:
identity = self.downsample(x)
out += identity
out = self.relu(out)
return out
3.5 空洞卷积具体写法
#定义replace_stride_with_dilation,如果采用空洞卷积，则设置stride为1，且将dilation参数*stride
def _make_layer(self, block, planes, blocks, stride=1, dilate=False):
...
previous_dilation = self.dilation
if dilation:
self.dilation *= stride
stride = 1 #如果使用空洞卷积，则stride重置为1
3.6 利用类+成员变量为参数赋值
#利用类+成员变量名来为单层赋值
nn.init.constant_(m.bn2.weight, 0) #注意这里的constant_