ResNet网络及代码

  残差网络是由来自 Microsoft Research 的 4 位学者提出的卷积神经网络，在 2015 年的 ImageNet 大规模视觉识别竞赛（ILSVRC）中获得了图像分类和物体识别的第一名，获得 COCO 数据集中目标检测第一名，图像分割第一名。残差网络的特点是容易优化，并且能够通过增加相当的深度来提高准确率。其内部的残差块使用了跳跃连接，缓解了在深度神经网络中增加深度带来的梯度消失问题。

  论文《Deep Residual Learning for Image Recognition》 论文链接

  当更深的网络能够开始收敛时，暴露了一个退化问题：随着网络深度的增加，准确率达到饱和然后迅速下降。意外的是，这种退化不是由过拟合引起的，并且在适当的深度模型上添加更多的层会导致更高的训练误差。论文提到更深的网络有更高的训练误差和测试误差，如下图所示，发现 56 层的训练误差（左）和测试误差（右）均比 20 层高。

  ResNet 的主要思想是在网络中增加了直连通道，允许原始输入信息直接传到后面的层中，如下图所示。传统的卷积层或全连接层在信息传递时或多或少会存在信息丢失问题，还会导致梯度消失/梯度爆炸，导致很深的网络无法训练。ResNet 在一定程度上解决了这个问题，通过直接将输入信息绕道传到输出，保护信息的完整性。

ResNet 残差学习模块

ResNet 亮点

（1）超深的网络结构(突破 1000 层)
（2）提出 residual 模块，解决退化问题
（3）使用 Batch Normalization 加速训练(丢弃 dropout)

ResNet 网络结构图

简单网络的基准（中间）主要受到 VGG 网络（左图）的哲学启发。

左：作为参考的 VGG-19 模型
中：具有 34 个参数层的简单网络
右：具有 34 个参数层的残差网络

ResNet网络表

ResNet 有不同的网络层数，包括18/34/50/101/152-layer
这五个层数 18/34/50/101/152-layer
相同点：
（1）卷积层 conv1 一样，均为 64 个 7*7 卷积核，步长 stride = 2；紧接着经过最大池化层 3*3 池化核，步长 stride = 2。
（2）conv5_x 后的 avepool，全连接层，softmax 一样。
不同点：
在 conv1 后的 maxpool 和 conv5_x 后的 avepool 之间，即 conv2_x ~ conv5_x 之间，不同层数的网络具有不同卷积核大小和数量。

代码

这里使用 pytorch 框架构建 ResNet 模型，训练花卉数据集
model.py

import torch.nn as nn
import torch

# 针对18/34层网络的conv2_x，conv3_x，conv4_x，conv5_x的系列卷积层
class BasicBlock(nn.Module):
    expansion = 1

    def __init__(self, in_channel, out_channel, stride=1, downsample=None):
        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:
            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

# 针对50/101/152层网络的conv2_x，conv3_x，conv4_x，conv5_x的系列卷积层
class Bottleneck(nn.Module):
    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)  # 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)

        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

# 网络ResNet
class ResNet(nn.Module):
    def __init__(self, block, blocks_num, num_classes=1000, include_top=True):
        # block有两种：BasicBlock针对18/34层网络，Bottleneck针对50/101/152层网络
        # blocks_num是一个列表，表示conv2_x，conv3_x，conv4_x，conv5_x分别对应的卷积层个数
        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], stride=1)   # 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')

    # 构建conv2_x，conv3_x，conv4_x，conv5_x卷积层
    def _make_layer(self, block, channel, block_num, stride=1):
        downsample = None  # 设定不是虚线,downsample不为None即是虚线
        # 网络结构中虚线路径的设定，只有18/34层网络的conv2_x不执行if语句
        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 = []
        # conv2_x，conv3_x，conv4_x，conv5_x的第一个卷积层
        # 第一层是虚线路径，传入downsample，因为两个block里面默认downsample = None
        layers.append(block(self.in_channel, channel, downsample=downsample, stride=stride))
        self.in_channel = channel * block.expansion
        # conv2_x，conv3_x，conv4_x，conv5_x每个系列的剩余卷积层，均为实线
        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)  # conv1
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)  # maxpool
        x = self.layer1(x)  # conv2_x
        x = self.layer2(x)  # conv3_x
        x = self.layer3(x)  # conv4_x
        x = self.layer4(x)  # conv5_x
        if self.include_top:
            x = self.avgpool(x)  # avgpool
            x = torch.flatten(x, 1)
            x = self.fc(x)  # fc

        return x


def resnet18(num_classes=1000, include_top=True):
    return ResNet(BasicBlock, [2, 2, 2, 2], num_classes=num_classes, include_top=include_top)
def resnet34(num_classes=1000, include_top=True):
    return ResNet(BasicBlock, [3, 4, 6, 3], num_classes=num_classes, include_top=include_top)

def resnet50(num_classes=1000, include_top=True):
    return ResNet(Bottleneck, [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)
def resnet152(num_classes=1000, include_top=True):
    return ResNet(Bottleneck, [3, 8, 36, 3], num_classes=num_classes, include_top=include_top)

train.py

import torch
import torch.nn as nn
from torchvision import transforms, datasets
import json
import matplotlib.pyplot as plt
import os
import torch.optim as optim
from model import resnet34, resnet101

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print(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(), "../..")) 
image_path = data_root + "/putty/resnet/data_set/junk_data/"  

train_dataset = datasets.ImageFolder(root=image_path+"train",
                                     transform=data_transform["train"])
train_num = len(train_dataset)

flower_list = train_dataset.class_to_idx
cla_dict = dict((val, key) for key, val in flower_list.items())
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
train_loader = torch.utils.data.DataLoader(train_dataset,
                                           batch_size=batch_size, shuffle=True,
                                           num_workers=0)

validate_dataset = datasets.ImageFolder(root=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=0)

net = resnet34()
# load pretrain weights
model_weight_path = "./resnet34-pre.pth"
missing_keys, unexpected_keys = net.load_state_dict(torch.load(model_weight_path), strict=False)
inchannel = net.fc.in_features
net.fc = nn.Linear(inchannel, 4)
net.to(device)

loss_function = nn.CrossEntropyLoss()
optimizer = optim.Adam(net.parameters(), lr=0.0001)

best_acc = 0.0
save_path = './resNet34.pth'
for epoch in range(2):
    # train
    net.train()
    running_loss = 0.0
    for step, data in enumerate(train_loader, start=0):
        images, labels = data
        optimizer.zero_grad()
        logits = net(images.to(device))
        loss = loss_function(logits, labels.to(device))
        loss.backward()
        optimizer.step()

        running_loss += loss.item()
        rate = (step+1)/len(train_loader)
        a = "*" * int(rate * 50)
        b = "." * int((1 - rate) * 50)
        print("\rtrain loss: {:^3.0f}%[{}->{}]{:.4f}".format(int(rate*100), a, b, loss), end="")
    print()

    # validate
    net.eval()
    acc = 0.0 
    with torch.no_grad():
        for val_data in validate_loader:
            val_images, val_labels = val_data
            outputs = net(val_images.to(device))
            predict_y = torch.max(outputs, dim=1)[1]
            acc += (predict_y == val_labels.to(device)).sum().item()
        val_accurate = acc / val_num
        if val_accurate > best_acc:
            best_acc = val_accurate
            torch.save(net.state_dict(), save_path)
        print('[epoch %d] train_loss: %.3f  test_accuracy: %.3f' %
              (epoch + 1, running_loss / step, val_accurate))

print('Finished Training')

predict.py

import torch
from model import resnet34
from PIL import Image
from torchvision import transforms
import matplotlib.pyplot as plt
import json

data_transform = transforms.Compose(
    [transforms.Resize(256),
     transforms.CenterCrop(224),
     transforms.ToTensor(),
     transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])

img = Image.open("data_set/junk_data/img_1.jpg")
plt.imshow(img)
img = data_transform(img)
img = torch.unsqueeze(img, dim=0)

try:
    json_file = open('./class_indices.json', 'r')
    class_indict = json.load(json_file)
except Exception as e:
    print(e)
    exit(-1)

model = resnet34(num_classes=4)
model_weight_path = "./resNet34.pth"
model.load_state_dict(torch.load(model_weight_path))
model.eval()
with torch.no_grad():
    output = torch.squeeze(model(img))
    predict = torch.softmax(output, dim=0)
    predict_cla = torch.argmax(predict).numpy()
print(class_indict[str(predict_cla)], predict[predict_cla].numpy())

name1 = class_indict[str(predict_cla)]
name2 =predict[predict_cla].numpy()
plt.title("This is %s. The accuracy is %s"%(name1, name2),color='red')
plt.show()