Resnet

DL DL Resnet

发布时间 : 2021-05-05 00:00

字数:3.1k 阅读 :

评论:

Resnet
1. 理论详解
2. 使用pytorch搭建ResNet并基于迁移学习训练

Resnet

理论详解

Resnet-34

梯度消失或梯度爆炸：BN解决
退化问题：残差结构解决

残差结构

同时输入 256-d(d代表维度) 数据，需要的参数 右边的残差结构 远远小于 左边的残差结构。

卷积核可以降维和升纬

看 34-layer !

网络结构中的虚线：

虚线的含义：调整参数结构，只有在层与层交界的地方才有。
- 以 2层到 3层为例：
  - 通过卷积核 1*1, 128 和 stride=2
  - 使得2层的输出结构与3层的结构保持一致：28*28, 128

更深的层：

参考博客: Resnet模型

第一层

18-layer 和 34-layer 第一层没有虚线

50-layer、101-layer、152-layer 需要虚线调整特征矩阵的深度，高宽不变

Batch Normalization

d 即 channels

迁移学习

使用pytorch搭建ResNet并基于迁移学习训练

18-layer 和 34-layer的残差结构

# 主分支卷积核个数是否发生变化
expansion = 1

在50-layer 、101-layer 、152-layer 中则 expansion = 4

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)

bias=False 表示不使用偏置（使用batch normalization时，不使用偏置）

downsample=None 表示主分支-实线不使用下采样，若不为None，则捷径-虚线使用下采样。

输出特征矩阵的 w-高 和 h-宽 :
- stride=1时：w,h 不变
- stride=2时：w,h 变为一半

identity: 通过下采样downsample得到捷径分支的输出

# 18-layer 和 34-layer的残差结构
class BasicBlock(nn.Module):
    # 卷积核个数变化的倍数
    expansion = 1

    ''' 初始化
        in_channel      输入特征矩阵的深度   
        out_channel     输出特征矩阵的深度   
        stride=1        步距为1对应实线   
        downsample=None 下采样-对应虚线的残差结构
    '''
    def __init__(self, in_channel, out_channel, stride=1, downsample=None, **kwargs):
        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):
        # 捷径分支的输出
        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

50-layer、101-layer、152-layer的残差结构

# 卷积核个数变化的倍数
expansion = 4

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)

identity: 通过下采样downsample得到捷径分支的输出

# 50-layer、101-layer、152-layer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
    """
    # 卷积核个数变化的倍数
    expansion = 4

    ''' 初始化
            in_channel      输入特征矩阵的深度   
            out_channel     输出特征矩阵的深度   
            stride=1        步距为1对应实线   
            downsample=None 下采样-对应虚线的残差结构
    '''
    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):
        # 捷径分支的输出
        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

ResNet框架结构

Model.py

import torch.nn as nn
import torch


# 18-layer 和 34-layer的残差结构
class BasicBlock(nn.Module):
    # 卷积核个数变化的倍数
    expansion = 1

    ''' 初始化
        in_channel      输入特征矩阵的深度   
        out_channel     输出特征矩阵的深度   
        stride=1        步距为1对应实线   
        downsample=None 下采样-对应虚线的残差结构
    '''
    def __init__(self, in_channel, out_channel, stride=1, downsample=None, **kwargs):
        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):
        # 捷径分支的输出
        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


# 50-layer、101-layer、152-layer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
    """
    # 卷积核个数变化的倍数
    expansion = 4

    ''' 初始化
            in_channel      输入特征矩阵的深度   
            out_channel     输出特征矩阵的深度   
            stride=1        步距为1对应实线   
            downsample=None 下采样-对应虚线的残差结构
    '''
    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):
        # 捷径分支的输出
        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):
    """
        block       对应残差结构
        blocks_num  对应所使用的的残差结构的数目形成的列表
        num_classes 分类数目，需要自己根据实际情况修改
        include_top=True    方便以后在resnet的基础上搭建更复杂的网络
    """
    def __init__(self,
                 block,
                 blocks_num,
                 num_classes=1000,
                 include_top=True,
                 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

        # 34-layer：7*7， 64， stride=2     w,h变为一半：112*112
        # self.in_channel 使用的卷积核个数
        # 3 代表 RGB 图像的深度
        self.conv1 = nn.Conv2d(3, self.in_channel, kernel_size=7, stride=2,
                               padding=3, bias=False)
        self.bn1 = nn.BatchNorm2d(self.in_channel)  # 输入即特征矩阵的输出深度
        self.relu = nn.ReLU(inplace=True)
        # 最大池化下采样后w,h变为一半：56*56
        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)
        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
        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,
                            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)

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, resnet50


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 = os.path.join(data_root, "data_set", "flower_data")  # 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)

    # {'daisy':0, 'dandelion':1, 'roses':2, 'sunflower':3, 'tulips':4}
    flower_list = train_dataset.class_to_idx
    cla_dict = dict((val, key) for key, val in flower_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=nw)

    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=nw)

    print("using {} images for training, {} images for validation.".format(train_num,
                                                                           val_num))
    # 实例化resnet34, 默认num_classes=1000， 可以根据自己需要指定分类类别
    num_classes = 6
    net = resnet34(num_classes)
    # net = resnet50(num_classes)
    # load pretrain weights 载入官方的权重预处理
    # download url: https://download.pytorch.org/models/resnet34-333f7ec4.pth
    model_weight_path = "./resnet34-pre.pth"
    # model_weight_path = "./resnet50-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))     # 载入模型权重
    # 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, 5)   # 5 表示分类类别个数 需要修改
    net.fc = nn.Linear(in_channel, num_classes)
    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()

predict.py

import os
import json

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

from model import resnet34, resnet50


def main():
    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

    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])])

    # load image
    img_path = "../1.jpg"
    assert os.path.exists(img_path), "file: '{}' dose not exist.".format(img_path)
    img = Image.open(img_path)
    plt.imshow(img)
    # [N, C, H, W]
    img = data_transform(img)
    # expand batch dimension
    img = torch.unsqueeze(img, dim=0)

    # read class_indict
    json_path = './class_indices.json'
    assert os.path.exists(json_path), "file: '{}' dose not exist.".format(json_path)

    json_file = open(json_path, "r")
    class_indict = json.load(json_file)

    # create model  需要修改 num_classes=5
    num_classes = 6
    model = resnet34(num_classes).to(device)
    # model = resnet50(num_classes).to(device)

    # load model weights
    weights_path = "./resNet34.pth"
    # weights_path = "./resNet50.pth"
    assert os.path.exists(weights_path), "file: '{}' dose not exist.".format(weights_path)
    model.load_state_dict(torch.load(weights_path, map_location=device))

    # prediction
    model.eval()
    with torch.no_grad():
        # predict class
        output = torch.squeeze(model(img.to(device))).cpu()
        predict = torch.softmax(output, dim=0)
        predict_cla = torch.argmax(predict).numpy()     # argmax最大值对应的索引，numpy再转为ndarray变量

    print_res = "class: {}   prob: {:.3}".format(class_indict[str(predict_cla)],
                                                 predict[predict_cla].numpy())
    plt.title(print_res)
    print(print_res)
    plt.show()


if __name__ == '__main__':
    main()

转载请注明来源，欢迎对文章中的引用来源进行考证，欢迎指出任何有错误或不够清晰的表达。可以在下面评论区评论，也可以邮件至 jungle8884@163.com