一、要点

ResNeXt是ResNet的小幅升级，更新了block

 左边（ResNet的block/50/101/152层）：

对于输入通道为256的特征矩阵，首先使用64个1×1的卷积核进行降维，再通过64个3×3的卷积核处理，再通过256个1×1的卷积核升维输出，将输出与输入进行相加，得到最终的输出。

使用右边的结构替代左边的结构：下面解释。

（一）论文中的性能参数指标

 （二）关于ResNet和ResNeXt在ImageNet上top-1 的错误率（计算量相同）

（三）组卷积 （Group Convolution）

当分组的个数与输入特征矩阵的channel是一致的，并且输入特征矩阵的channel也和输出特征矩阵的channel一致的话，就相当于对我们输入特征矩阵的每一个channel分配了一个channel为1的卷积核进行卷积。即DW卷积。

（四）ResNeXt的block结构

（c）（最简形式）:输入通道为256维，首先通过128个1×1的卷积核降维处理，再通过group卷积（卷积核3×3，group数为32），得到的特征矩阵的通道是128维，再通过256个1×1的卷积核升维得到输出。再将输出和输入的特征矩阵进行相加得到最终的输出。

（b）和（c）等价：

 （a）和（b）等价：

 举例：假设path为2，对每个path采用1×1的卷积核来进行卷积

（五） 网络结构

二、使用pytorch搭建

代码是包括ResNet和ResNeXt的

import torch.nn as nn
import torch

# 18层/34层 对应的残差结构（既要有实线残差结构的功能，又要有虚线残差结构的功能）
class BasicBlock(nn.Module):
    expansion = 1  #残差结构的主分支卷积核的个数有无发生变化

    def __init__(self, in_channel, out_channel, stride=1, downsample=None, **kwargs): # downsample对应残差结构的虚线（shortcut分支）
        super(BasicBlock, self).__init__()
        # stride=1,对应着实线的残差结构，因为并没有改变输入特征矩阵的高和宽
        # output = (input -3 + 2 * 1) / 1 + 1 = input
        # stride=2,对应着虚线的残差结构，在第一个卷积层需要将特征矩阵的高和宽缩减为原来的一半
        # output = (input -3 + 2 * 1) / 2 + 1 = input / 2 + 0.5 = input / 2 （向下取整）
        self.conv1 = nn.Conv2d(in_channels=in_channel, out_channels=out_channel,
                               kernel_size=3, stride=stride, padding=1, bias=False)  # bias=False,不使用偏置项，因为下面用到BatchNormalization
        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=False)
        self.bn2 = nn.BatchNorm2d(out_channel)
        self.downsample = downsample

    def forward(self, x):
        identity = x # shorcut上的输出值
        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 += identity
        out = self.relu(out)

        return out

# 50层/101层/152层 对应的残差结构
class Bottleneck(nn.Module):
    """
    注意：原论文中，在虚线残差结构的主分支上，第一个1x1卷积层的步距是2，第二个3x3卷积层步距是1。
    但在pytorch官方实现过程中是第一个1x1卷积层的步距是1，第二个3x3卷积层步距是2，
    这么做的好处是能够在top1上提升大概0.5%的准确率。
    可参考Resnet v1.5 https://ngc.nvidia.com/catalog/model-scripts/nvidia:resnet_50_v1_5_for_pytorch
    """
    expansion = 4  #每个残差结构的最后一层卷积核的个数都是前两层的4倍

    def __init__(self, in_channel, out_channel, stride=1, downsample=None,
                 groups=1, width_per_group=64):
        super(Bottleneck, self).__init__()

        width = int(out_channel * (width_per_group / 64.)) * groups

        self.conv1 = nn.Conv2d(in_channels=in_channel, out_channels=width,
                               kernel_size=1, stride=1, bias=False)  # squeeze channels
        self.bn1 = nn.BatchNorm2d(width)
        # -----------------------------------------
        self.conv2 = nn.Conv2d(in_channels=width, out_channels=width, groups=groups,
                               kernel_size=3, stride=stride, bias=False, padding=1)
        self.bn2 = nn.BatchNorm2d(width)
        # -----------------------------------------
        self.conv3 = nn.Conv2d(in_channels=width, out_channels=out_channel*self.expansion,
                               kernel_size=1, stride=1, bias=False)  # unsqueeze channels
        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


class ResNet(nn.Module):

    def __init__(self,
                 block, # 对应18/34层或者50/101/152层的残差结构
                 blocks_num, # 残差结构的个数，是个列表，以34层为例，就是[3,4,6,3]
                 num_classes=1000, # 训练集的分类个数
                 include_top=True, # 方便以后在ResNet网络上去搭建更复杂的网络
                 groups=1,
                 width_per_group=64):
        super(ResNet, self).__init__()
        self.include_top = include_top
        self.in_channel = 64

        self.groups = groups
        self.width_per_group = width_per_group

        self.conv1 = nn.Conv2d(3, self.in_channel, kernel_size=7, stride=2,
                               padding=3, bias=False)# 为了让特征矩阵的宽和高缩减为原来的一半，所以这里padding=3
        self.bn1 = nn.BatchNorm2d(self.in_channel)
        self.relu = nn.ReLU(inplace=True)
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)# 为了让特征矩阵的宽和高缩减为原来的一半，所以这里padding=1
        self.layer1 = self._make_layer(block, 64, blocks_num[0]) # conv2_x 对于50/101/152层来说，第一个残差结构的第一层只改变特征矩阵的深度，没有改变宽高，所以没有传入stride，默认为1
        self.layer2 = self._make_layer(block, 128, blocks_num[1], stride=2) # conv3_x
        self.layer3 = self._make_layer(block, 256, blocks_num[2], stride=2) # conv4_x
        self.layer4 = self._make_layer(block, 512, blocks_num[3], stride=2) # conv5_x
        if self.include_top:
            self.avgpool = nn.AdaptiveAvgPool2d((1, 1))  # output size = (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')

    def _make_layer(self, block, channel, block_num, stride=1):
        downsample = None
        # 对于18/34层第一个残差结构，跳过这句
        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 = []
        # 残差结构的第一层（虚线残差结构）
        layers.append(block(self.in_channel,
                            channel,
                            downsample=downsample,
                            stride=stride,
                            groups=self.groups,
                            width_per_group=self.width_per_group))
        self.in_channel = channel * block.expansion

        for _ in range(1, block_num):
            layers.append(block(self.in_channel,
                                channel,
                                groups=self.groups,
                                width_per_group=self.width_per_group))

        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)

        if self.include_top:
            x = self.avgpool(x)
            x = torch.flatten(x, 1)
            x = self.fc(x)

        return x


def resnet34(num_classes=1000, include_top=True):
    # https://download.pytorch.org/models/resnet34-333f7ec4.pth
    return ResNet(BasicBlock, [3, 4, 6, 3], num_classes=num_classes, include_top=include_top)


def resnet50(num_classes=1000, include_top=True):
    # https://download.pytorch.org/models/resnet50-19c8e357.pth
    return ResNet(Bottleneck, [3, 4, 6, 3], num_classes=num_classes, include_top=include_top)


def resnet101(num_classes=1000, include_top=True):
    # https://download.pytorch.org/models/resnet101-5d3b4d8f.pth
    return ResNet(Bottleneck, [3, 4, 23, 3], num_classes=num_classes, include_top=include_top)


def resnext50_32x4d(num_classes=1000, include_top=True):
    # https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth
    groups = 32
    width_per_group = 4
    return ResNet(Bottleneck, [3, 4, 6, 3],
                  num_classes=num_classes,
                  include_top=include_top,
                  groups=groups,
                  width_per_group=width_per_group)


def resnext101_32x8d(num_classes=1000, include_top=True):
    # https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth
    groups = 32
    width_per_group = 8
    return ResNet(Bottleneck, [3, 4, 23, 3],
                  num_classes=num_classes,
                  include_top=include_top,
                  groups=groups,
                  width_per_group=width_per_group)