第3周学习：ResNet+ResNeXt

一.ResNet

1.ResNet

ResNet在2015年由微软实验室提出，斩获当年lmageNet竞赛中分类任务第一名，目标检测第一名。获得coco数据集中目标检测第一名，图像分割第一名。

网络中的亮点:

超深的网络结构(突破1000层)；

提出residual模块；

使用Batch Normalization加速训练(丢弃dropout)。
 

简单堆叠的网络结构（卷积层和池化层）并不会层数越多效果越好。

随着网络层数加深，会出现两个问题越来越明显：

梯度消失或者梯度爆炸；

退化问题。

 

两种residual结构：

主分支与shortcut的输出特征矩阵shape必须相同。右图的参数量要小于左图。

 

右侧shortcut分支加入1x1卷积层，维度层数相同（64->128），使得残差结构输入和输出大小一致。左侧主分支采用stride=2的3x3卷，128的卷积，使减半（56x56->28x28）。

2.Batch Normalization

Batch Normalization是google团队在2015年论文《Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift》提出的。通过该方法能够加速网络的收敛并提升准确率。在网上虽然已经有很多相关文章，但基本都是摆上论文中的公式泛泛而谈，bn真正是如何运作的很少有提及。

我们在图像预处理过程中通常会对图像进行标准化处理，这样能够加速网络的收敛，如下图所示，对于Conv1来说输入的就是满足某一分布的特征矩阵，但对于Conv2而言输入的feature map就不一定满足某一分布规律了（注意这里所说满足某一分布规律并不是指某一个feature map的数据要满足分布规律，理论上是指整个训练样本集所对应feature map的数据要满足分布规律）。而我们Batch Normalization的目的就是使我们的feature map满足均值为0，方差为1的分布规律。

 原文提供了更加详细的计算公式：

3.迁移学习
 

优势：
能够快速的训练出一个理想的结果。
当数据集较小时也能训练出理想的效果。

需要注意：在使用其他预训练模型参数时，要注意其预处理方式。

常见的迁移学习方式：
载入权重后训练所有参数
载入权重后只训练最后几层参数
载入权重后在原网络基础上再添加一层全连接层，仅训练最后一个全连接层

一.ResNetXt

一种对 ResNet的升级改进。

分组卷积：

block模块：

对比如下图：

 

比如conv2对应的中间主卷积：C=32指的是每个block使用的 group数为32 ，4代表每一个group运用4个卷积核，输出为128。

大致框架几乎一样，将block进行重复的堆叠最后接上池化层全连接层。

三.代码练习

1.Lenet：

参考了高老师的博客LeNet 网络进行猫狗大战 - 高峰OUC - 博客园 (cnblogs.com)

定义网络：

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(3, 6, 5)
        self.conv2 = nn.Conv2d(6, 16, 5)
        self.conv3 = nn.Conv2d(16, 32, 4)
        self.conv4 = nn.Conv2d(32, 32, 4)
        self.conv5 = nn.Conv2d(32, 32, 5)
        self.pool = nn.MaxPool2d(2, 2)
        self.fc1 = nn.Linear(32, 16)
        self.fc2 = nn.Linear(16, 2)

    def forward(self, x):
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = self.pool(F.relu(self.conv3(x)))
        x = self.pool(F.relu(self.conv4(x)))
        x = F.relu(self.conv5(x))
        x = x.view(-1, 32)
        x = F.relu(self.fc1(x))
        x = F.softmax(self.fc2(x), dim=1)
        return x

 上传得到结果：

 

2.ResNet

import torch.nn as nn
import torch


# ResNet18/34的残差结构，用的是2个3x3的卷积
class BasicBlock(nn.Module):
    expansion = 1  # 残差结构中，主分支的卷积核个数是否发生变化，不变则为1

    def __init__(self, in_channel, out_channel, stride=1, downsample=None):  # downsample对应虚线残差结构
        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=False)
        self.bn2 = nn.BatchNorm2d(out_channel)
        self.downsample = downsample

    def forward(self, x):
        identity = x
        if self.downsample is not None:  # 虚线残差结构，需要下采样
            identity = self.downsample(x)  # 捷径分支 short cut

        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

# ResNet50/101/152的残差结构，用的是1x1+3x3+1x1的卷积
class Bottleneck(nn.Module):
    expansion = 4  # 残差结构中第三层卷积核个数是第一/二层卷积核个数的4倍

    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)  # squeeze channels
        self.bn1 = nn.BatchNorm2d(out_channel)
        # -----------------------------------------
        self.conv2 = nn.Conv2d(in_channels=out_channel, out_channels=out_channel,
                               kernel_size=3, stride=stride, bias=False, padding=1)
        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)  # 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)  # 捷径分支 short cut

        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):
    # block = BasicBlock or Bottleneck
    # block_num为残差结构中conv2_x~conv5_x中残差块个数，是一个列表
    def __init__(self, block, blocks_num, num_classes=1000, include_top=True):
        super(ResNet, self).__init__()
        self.include_top = include_top
        self.in_channel = 64

        self.conv1 = nn.Conv2d(3, self.in_channel, kernel_size=7, stride=2,
                               padding=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])             # conv2_x
        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')

    # channel为残差结构中第一层卷积核个数
    def _make_layer(self, block, channel, block_num, stride=1):
        downsample = None

        # ResNet50/101/152的残差结构，block.expansion=4
        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)

        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):
    return ResNet(BasicBlock, [3, 4, 6, 3], num_classes=num_classes, include_top=include_top)


def resnet101(num_classes=1000, include_top=True):
    return ResNet(Bottleneck, [3, 4, 23, 3], num_classes=num_classes, include_top=include_top)

train.py

import os
import json

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import transforms, datasets
from tqdm import tqdm

from model import resnet34


def main():
    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
    print("using {} device.".format(device))

    data_transform = {
        "train": transforms.Compose([transforms.RandomResizedCrop(224),
                                     transforms.RandomHorizontalFlip(),
                                     transforms.ToTensor(),
                                     transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]),
        "val": transforms.Compose([transforms.Resize(256),
                                   transforms.CenterCrop(224),
                                   transforms.ToTensor(),
                                   transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])}

    data_root = os.path.abspath(os.path.join(os.getcwd(), "../.."))  # get data root path
    image_path = "cat_dog/"  # flower data set path
    assert os.path.exists(image_path), "{} path does not exist.".format(image_path)
    train_dataset = datasets.ImageFolder(root=os.path.join(image_path, "train"),
                                         transform=data_transform["train"])
    train_num = len(train_dataset)

    # {'cat':0, 'dog':1}
    animal_list = train_dataset.class_to_idx
    cla_dict = dict((val, key) for key, val in animal_list.items())
    # write dict into json file
    json_str = json.dumps(cla_dict, indent=4)
    with open('class_indices.json', 'w') as json_file:
        json_file.write(json_str)

    batch_size = 16
    nw = min([os.cpu_count(), batch_size if batch_size > 1 else 0, 8])  # number of workers
    print('Using {} dataloader workers every process'.format(nw))

    train_loader = torch.utils.data.DataLoader(train_dataset,
                                               batch_size=batch_size, shuffle=True,
                                               num_workers=2)

    validate_dataset = datasets.ImageFolder(root=os.path.join(image_path, "val"),
                                            transform=data_transform["val"])
    val_num = len(validate_dataset)
    validate_loader = torch.utils.data.DataLoader(validate_dataset,
                                                  batch_size=batch_size, shuffle=False,
                                                  num_workers=2)

    print("using {} images for training, {} images for validation.".format(train_num,
                                                                           val_num))
    
    net = resnet34()
    # load pretrain weights
    # download url: https://download.pytorch.org/models/resnet34-333f7ec4.pth
    model_weight_path = "./resnet34-pre.pth"
    assert os.path.exists(model_weight_path), "file {} does not exist.".format(model_weight_path)
    net.load_state_dict(torch.load(model_weight_path, map_location=device))
    # for param in net.parameters():
    #     param.requires_grad = False

    # change fc layer structure
    in_channel = net.fc.in_features
    net.fc = nn.Linear(in_channel, 2)
    net.to(device)

    # define loss function
    loss_function = nn.CrossEntropyLoss()

    # construct an optimizer
    params = [p for p in net.parameters() if p.requires_grad]
    optimizer = optim.Adam(params, lr=0.0001)

    epochs = 3
    best_acc = 0.0
    save_path = './resNet34.pth'
    train_steps = len(train_loader)
    for epoch in range(epochs):
        # train
        net.train()
        running_loss = 0.0
        train_bar = tqdm(train_loader)
        for step, data in enumerate(train_bar):
            images, labels = data
            optimizer.zero_grad()
            logits = net(images.to(device))
            loss = loss_function(logits, labels.to(device))
            loss.backward()
            optimizer.step()

            # print statistics
            running_loss += loss.item()

            train_bar.desc = "train epoch[{}/{}] loss:{:.3f}".format(epoch + 1,
                                                                     epochs,
                                                                     loss)

        # validate
        net.eval()
        acc = 0.0  # accumulate accurate number / epoch
        with torch.no_grad():
            val_bar = tqdm(validate_loader)
            for val_data in val_bar:
                val_images, val_labels = val_data
                outputs = net(val_images.to(device))
                # loss = loss_function(outputs, test_labels)
                predict_y = torch.max(outputs, dim=1)[1]
                acc += torch.eq(predict_y, val_labels.to(device)).sum().item()

                val_bar.desc = "valid epoch[{}/{}]".format(epoch + 1,
                                                           epochs)

        val_accurate = acc / val_num
        print('[epoch %d] train_loss: %.3f  val_accuracy: %.3f' %
              (epoch + 1, running_loss / train_steps, val_accurate))

        if val_accurate > best_acc:
            best_acc = val_accurate
            torch.save(net.state_dict(), save_path)

    print('Finished Training')


if __name__ == '__main__':
    main()

但训练三轮过后，没有得到预期效果，效果很差。待gpu恢复后继续尝试。

四.问题与思考

1、Residual learning

 在此模型中，由于x直接连接到下一层，没有任何参数，即网络学习到的是F(x)。

2、Batch Normailization 的原理

BN层的计算流程是：

1.计算样本均值。

2.计算样本方差。

3.样本数据标准化处理。

4.进行平移和缩放处理。引入了γ和β两个参数。来训练γ和β两个参数。引入了这个可学习重构参数γ、β，让我们的网络可以学习恢复出原始网络所要学习的特征分布。

3、为什么分组卷积可以提升准确率？即然分组卷积可以提升准确率，同时还能降低计算量，分数数量尽量多不行吗？

提升准确率原因：减少运算量和参数量，相同输入输出大小的情况下，减少为原来的 1g ；隔绝不同组的信息交换。

尽量多不可以，因为组和组的联系是需要时间以及其他资源分配的，会出现信息的互相隔离。