动手学深度学习PyTorch版--Task8--图像分类案例1；图像分类案例2

一.案例1:Kaggle上的图像分类（CIFAR-10)

现在，我们将运用在前面几节中学到的知识来参加Kaggle竞赛，该竞赛解决了CIFAR-10图像分类问题。比赛网址是https://www.kaggle.com/c/cifar-10

# 本节的网络需要较长的训练时间
# 可以在Kaggle访问：
# https://www.kaggle.com/boyuai/boyu-d2l-image-classification-cifar-10
import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim
import torchvision
import torchvision.transforms as transforms
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
import os
import time
print("PyTorch Version: ",torch.__version__)
'''
PyTorch Version:  1.3.0
'''

首先我们设置随机数种子，确保在相同参数下，可以得到一致的运行结果。

random.seed(0)
torch.manual_seed(0)
torch.cuda.manual_seed(0)

1.获取和组织数据集

比赛数据分为训练集和测试集。训练集包含 50,000 图片。测试集包含 300,000 图片。两个数据集中的图像格式均为PNG，高度和宽度均为32像素，并具有三个颜色通道（RGB）。图像涵盖10个类别：飞机，汽车，鸟类，猫，鹿，狗，青蛙，马，船和卡车。 为了更容易上手，我们提供了上述数据集的小样本。“ train_tiny.zip”包含 80 训练样本，而“ test_tiny.zip”包含100个测试样本。它们的解压文件夹名称分别是“ train_tiny”和“ test_tiny”。

2.图像增强

data_transform = transforms.Compose([
    transforms.Resize(40),
    transforms.RandomHorizontalFlip(),
    transforms.RandomCrop(32),
    transforms.ToTensor()
])
trainset = torchvision.datasets.ImageFolder(root='/home/kesci/input/CIFAR102891/cifar-10/train'
                                            , transform=data_transform)
# 查看图像形状
trainset[0][0].shape
# 计算均值
data = [d[0].data.cpu().numpy() for d in trainset]
np.mean(data)
# 计算标准差
np.std(data)
'''
torch.Size([3, 32, 32])
0.4676536
0.23926772
'''

# 图像增强
transform_train = transforms.Compose([
    transforms.RandomCrop(32, padding=4),  #先四周填充0，再把图像随机裁剪成32*32
    transforms.RandomHorizontalFlip(),  #图像一半的概率翻转，一半的概率不翻转
    transforms.ToTensor(),
    transforms.Normalize((0.4731, 0.4822, 0.4465), (0.2212, 0.1994, 0.2010)), #R,G,B每层的归一化用到的均值和方差
])

transform_test = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize((0.4731, 0.4822, 0.4465), (0.2212, 0.1994, 0.2010)),
])

3.导入数据集

train_dir = '/home/kesci/input/CIFAR102891/cifar-10/train'
test_dir = '/home/kesci/input/CIFAR102891/cifar-10/test'

trainset = torchvision.datasets.ImageFolder(root=train_dir, transform=transform_train)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=256, shuffle=True)

testset = torchvision.datasets.ImageFolder(root=test_dir, transform=transform_test)
testloader = torch.utils.data.DataLoader(testset, batch_size=256, shuffle=False)

classes = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'forg', 'horse', 'ship', 'truck']

4.定义模型

ResNet-18网络结构：ResNet全名Residual Network残差网络。Kaiming He 的《Deep Residual Learning for Image Recognition》获得了CVPR最佳论文。他提出的深度残差网络在2015年可以说是洗刷了图像方面的各大比赛，以绝对优势取得了多个比赛的冠军。而且它在保证网络精度的前提下，将网络的深度达到了152层，后来又进一步加到1000的深度。

class ResidualBlock(nn.Module):   # 我们定义网络时一般是继承的torch.nn.Module创建新的子类

    def __init__(self, inchannel, outchannel, stride=1):
        super(ResidualBlock, self).__init__()
        #torch.nn.Sequential是一个Sequential容器，模块将按照构造函数中传递的顺序添加到模块中。
        self.left = nn.Sequential(
            nn.Conv2d(inchannel, outchannel, kernel_size=3, stride=stride, padding=1, bias=False), 
            # 添加第一个卷积层,调用了nn里面的Conv2d（）
            nn.BatchNorm2d(outchannel), # 进行数据的归一化处理
            nn.ReLU(inplace=True), # 修正线性单元，是一种人工神经网络中常用的激活函数
            nn.Conv2d(outchannel, outchannel, kernel_size=3, stride=1, padding=1, bias=False),
            nn.BatchNorm2d(outchannel)
        )
        self.shortcut = nn.Sequential() 
        if stride != 1 or inchannel != outchannel:
            self.shortcut = nn.Sequential(
                nn.Conv2d(inchannel, outchannel, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(outchannel)
            )
        #  便于之后的联合,要判断Y = self.left(X)的形状是否与X相同

    def forward(self, x): # 将两个模块的特征进行结合，并使用ReLU激活函数得到最终的特征。
        out = self.left(x)
        out += self.shortcut(x)
        out = F.relu(out)
        return out

class ResNet(nn.Module):
    def __init__(self, ResidualBlock, num_classes=10):
        super(ResNet, self).__init__()
        self.inchannel = 64
        self.conv1 = nn.Sequential( # 用3个3x3的卷积核代替7x7的卷积核，减少模型参数
            nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1, bias=False),
            nn.BatchNorm2d(64),
            nn.ReLU(),
        ) 
        self.layer1 = self.make_layer(ResidualBlock, 64,  2, stride=1)
        self.layer2 = self.make_layer(ResidualBlock, 128, 2, stride=2)
        self.layer3 = self.make_layer(ResidualBlock, 256, 2, stride=2)
        self.layer4 = self.make_layer(ResidualBlock, 512, 2, stride=2)
        self.fc = nn.Linear(512, num_classes)

    def make_layer(self, block, channels, num_blocks, stride):
        strides = [stride] + [1] * (num_blocks - 1)   #第一个ResidualBlock的步幅由make_layer的函数参数stride指定
        # ，后续的num_blocks-1个ResidualBlock步幅是1
        layers = []
        for stride in strides:
            layers.append(block(self.inchannel, channels, stride))
            self.inchannel = channels
        return nn.Sequential(*layers)

    def forward(self, x):
        out = self.conv1(x)
        out = self.layer1(out)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = F.avg_pool2d(out, 4)
        out = out.view(out.size(0), -1)
        out = self.fc(out)
        return out


def ResNet18():
    return ResNet(ResidualBlock)

5.训练和测试

# 定义是否使用GPU
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# 超参数设置
EPOCH = 20   #遍历数据集次数
pre_epoch = 0  # 定义已经遍历数据集的次数
LR = 0.1        #学习率

# 模型定义-ResNet
net = ResNet18().to(device)

# 定义损失函数和优化方式
criterion = nn.CrossEntropyLoss()  #损失函数为交叉熵，多用于多分类问题
optimizer = optim.SGD(net.parameters(), lr=LR, momentum=0.9, weight_decay=5e-4) 
#优化方式为mini-batch momentum-SGD，并采用L2正则化（权重衰减）

# 训练
if __name__ == "__main__":
    print("Start Training, Resnet-18!")
    num_iters = 0
    for epoch in range(pre_epoch, EPOCH):
        print('\nEpoch: %d' % (epoch + 1))
        net.train()
        sum_loss = 0.0
        correct = 0.0
        total = 0
        for i, data in enumerate(trainloader, 0): 
            #用于将一个可遍历的数据对象(如列表、元组或字符串)组合为一个索引序列，同时列出数据和数据下标，
            #下标起始位置为0，返回 enumerate(枚举) 对象。
            
            num_iters += 1
            inputs, labels = data
            inputs, labels = inputs.to(device), labels.to(device)
            optimizer.zero_grad()  # 清空梯度

            # forward + backward
            outputs = net(inputs)
            loss = criterion(outputs, labels)
            loss.backward()
            optimizer.step()

            sum_loss += loss.item() * labels.size(0)
            _, predicted = torch.max(outputs, 1) #选出每一列中最大的值作为预测结果
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
            # 每20个batch打印一次loss和准确率
            if (i + 1) % 20 == 0:
                print('[epoch:%d, iter:%d] Loss: %.03f | Acc: %.3f%% '
                        % (epoch + 1, num_iters, sum_loss / (i + 1), 100. * correct / total))

    print("Training Finished, TotalEPOCH=%d" % EPOCH)

二.案例2:Kaggle上的狗品种识别（ImageNet Dogs）

在本节中，我们将解决Kaggle竞赛中的犬种识别挑战，比赛的网址是https://www.kaggle.com/c/dog-breed-identification 在这项比赛中，我们尝试确定120种不同的狗。该比赛中使用的数据集实际上是著名的ImageNet数据集的子集。

# 在本节notebook中，使用后续设置的参数在完整训练集上训练模型，大致需要40-50分钟
# 请大家合理安排GPU时长，尽量只在训练时切换到GPU资源
# 也可以在Kaggle上访问本节notebook：
# https://www.kaggle.com/boyuai/boyu-d2l-dog-breed-identification-imagenet-dogs
import torch
import torch.nn as nn
import torch.optim as optim
import torchvision
import torchvision.transforms as transforms
import torchvision.models as models
import os
import shutil
import time
import pandas as pd
import random

# 设置随机数种子
random.seed(0)
torch.manual_seed(0)
torch.cuda.manual_seed(0)

1.整理数据集

我们可以从比赛网址上下载数据集，其目录结构为：

| Dog Breed Identification
| train
| | 000bec180eb18c7604dcecc8fe0dba07.jpg
| | 00a338a92e4e7bf543340dc849230e75.jpg
| | …
| test
| | 00a3edd22dc7859c487a64777fc8d093.jpg
| | 00a6892e5c7f92c1f465e213fd904582.jpg
| | …
| labels.csv
| sample_submission.csv

train和test目录下分别是训练集和测试集的图像，训练集包含10,222张图像，测试集包含10,357张图像，图像格式都是JPEG，每张图像的文件名是一个唯一的id。labels.csv包含训练集图像的标签，文件包含10,222行，每行包含两列，第一列是图像id，第二列是狗的类别。狗的类别一共有120种。

我们希望对数据进行整理，方便后续的读取，我们的主要目标是：

从训练集中划分出验证数据集，用于调整超参数。划分之后，数据集应该包含4个部分：划分后的训练集、划分后的验证集、完整训练集、完整测试集
对于4个部分，建立4个文件夹：train, valid, train_valid, test。在上述文件夹中，对每个类别都建立一个文件夹，在其中存放属于该类别的图像（在代码中 取img_label作为每个类别的文件夹）。前三个部分的标签已知，所以各有120个子文件夹，而测试集的标签未知，所以仅建立一个名为unknown的子文件夹，存放所有测试数据。
我们希望整理后的数据集目录结构为：

| train_valid_test
| train
| | affenpinscher
| | | 00ca18751837cd6a22813f8e221f7819.jpg
| | | …
| | afghan_hound
| | | 0a4f1e17d720cdff35814651402b7cf4.jpg
| | | …
| | …
| valid
| | affenpinscher
| | | 56af8255b46eb1fa5722f37729525405.jpg
| | | …
| | afghan_hound
| | | 0df400016a7e7ab4abff824bf2743f02.jpg
| | | …
| | …
| train_valid
| | affenpinscher
| | | 00ca18751837cd6a22813f8e221f7819.jpg
| | | …
| | afghan_hound
| | | 0a4f1e17d720cdff35814651402b7cf4.jpg
| | | …
| | …
| test
| | unknown
| | | 00a3edd22dc7859c487a64777fc8d093.jpg
| | | …

data_dir = '/home/kesci/input/Kaggle_Dog6357/dog-breed-identification'  # 数据集目录
label_file, train_dir, test_dir = 'labels.csv', 'train', 'test'  # data_dir中的文件夹、文件
new_data_dir = './train_valid_test'  # 整理之后的数据存放的目录
valid_ratio = 0.1  # 验证集所占比例

def mkdir_if_not_exist(path):
    # 若目录path不存在，则创建目录
    if not os.path.exists(os.path.join(*path)):
        os.makedirs(os.path.join(*path))
        
def reorg_dog_data(data_dir, label_file, train_dir, test_dir, new_data_dir, valid_ratio):
    # 读取训练数据标签
    labels = pd.read_csv(os.path.join(data_dir, label_file))
    id2label = {
     Id: label for Id, label in labels.values}  # (key: value): (id: label)

    # 随机打乱训练数据
    train_files = os.listdir(os.path.join(data_dir, train_dir))
    random.shuffle(train_files)    

    # 原训练集
    valid_ds_size = int(len(train_files) * valid_ratio)  # 验证集大小
    for i, file in enumerate(train_files):
        img_id = file.split('.')[0]  # file是形式为id.jpg的字符串
        img_label = id2label[img_id]
        if i < valid_ds_size:
        	# 取img_label作为每个类别的文件夹
            mkdir_if_not_exist([new_data_dir, 'valid', img_label])
            shutil.copy(os.path.join(data_dir, train_dir, file),
                        os.path.join(new_data_dir, 'valid', img_label))
        else:
            mkdir_if_not_exist([new_data_dir, 'train', img_label])
            shutil.copy(os.path.join(data_dir, train_dir, file),
                        os.path.join(new_data_dir, 'train', img_label))
        mkdir_if_not_exist([new_data_dir, 'train_valid', img_label])
        shutil.copy(os.path.join(data_dir, train_dir, file),
                    os.path.join(new_data_dir, 'train_valid', img_label))

    # 测试集
    mkdir_if_not_exist([new_data_dir, 'test', 'unknown'])
    for test_file in os.listdir(os.path.join(data_dir, test_dir)):
        shutil.copy(os.path.join(data_dir, test_dir, test_file),
                    os.path.join(new_data_dir, 'test', 'unknown'))

reorg_dog_data(data_dir, label_file, train_dir, test_dir, new_data_dir, valid_ratio)

2.图像增强

transform_train = transforms.Compose([
    # 随机对图像裁剪出面积为原图像面积0.08~1倍、且高和宽之比在3/4~4/3的图像，再放缩为高和宽均为224像素的新图像
    transforms.RandomResizedCrop(224, scale=(0.08, 1.0),  
                                 ratio=(3.0/4.0, 4.0/3.0)),
    # 以0.5的概率随机水平翻转
    transforms.RandomHorizontalFlip(),
    # 随机更改亮度、对比度和饱和度
    transforms.ColorJitter(brightness=0.4, contrast=0.4, saturation=0.4),
    transforms.ToTensor(),
    # 对各个通道做标准化，(0.485, 0.456, 0.406)和(0.229, 0.224, 0.225)是在ImageNet上计算得的各通道均值与方差
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])  # ImageNet上的均值和方差
])

# 在测试集上的图像增强只做确定性的操作
transform_test = transforms.Compose([
    transforms.Resize(256),
    # 将图像中央的高和宽均为224的正方形区域裁剪出来
    transforms.CenterCrop(224),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

3.读取数据

# new_data_dir目录下有train, valid, train_valid, test四个目录
# 这四个目录中，每个子目录表示一种类别，目录中是属于该类别的所有图像
train_ds = torchvision.datasets.ImageFolder(root=os.path.join(new_data_dir, 'train'),
                                            transform=transform_train)
valid_ds = torchvision.datasets.ImageFolder(root=os.path.join(new_data_dir, 'valid'),
                                            transform=transform_test)
train_valid_ds = torchvision.datasets.ImageFolder(root=os.path.join(new_data_dir, 'train_valid'),
                                            transform=transform_train)
test_ds = torchvision.datasets.ImageFolder(root=os.path.join(new_data_dir, 'test'),
                                            transform=transform_test)

batch_size = 128
train_iter = torch.utils.data.DataLoader(train_ds, batch_size=batch_size, shuffle=True)
valid_iter = torch.utils.data.DataLoader(valid_ds, batch_size=batch_size, shuffle=True)
train_valid_iter = torch.utils.data.DataLoader(train_valid_ds, batch_size=batch_size, shuffle=True)
test_iter = torch.utils.data.DataLoader(test_ds, batch_size=batch_size, shuffle=False)  # shuffle=False

4.定义模型

这个比赛的数据属于ImageNet数据集的子集，我们使用微调的方法，选用在ImageNet完整数据集上预训练的模型来抽取图像特征，以作为自定义小规模输出网络的输入。

此处我们使用与训练的ResNet-34模型，直接复用预训练模型在输出层的输入，即抽取的特征，然后我们重新定义输出层，本次我们仅对重定义的输出层的参数进行训练（requires_grad = False 冻结参数），而对于用于抽取特征的部分，我们保留预训练模型的参数。

def get_net(device):
    finetune_net = models.resnet34(pretrained=False)  # 预训练的resnet34网络
    finetune_net.load_state_dict(torch.load('/home/kesci/input/resnet347742/resnet34-333f7ec4.pth'))
    for param in finetune_net.parameters():  # 冻结参数
        param.requires_grad = False
    # 原finetune_net.fc是一个输入单元数为512，输出单元数为1000的全连接层
    # 替换掉原finetune_net.fc，新finetuen_net.fc中的模型参数会记录梯度
    finetune_net.fc = nn.Sequential(
        nn.Linear(in_features=512, out_features=256),
        nn.ReLU(),
        nn.Linear(in_features=256, out_features=120)  # 120是输出类别数
    )
    return finetune_net

5.定义训练函数

def evaluate_loss_acc(data_iter, net, device):
    # 计算data_iter上的平均损失与准确率
    loss = nn.CrossEntropyLoss()
    is_training = net.training  # Bool net是否处于train模式
    net.eval()
    l_sum, acc_sum, n = 0, 0, 0
    with torch.no_grad():
        for X, y in data_iter:
            X, y = X.to(device), y.to(device)
            y_hat = net(X)
            l = loss(y_hat, y)
            l_sum += l.item() * y.shape[0]
            acc_sum += (y_hat.argmax(dim=1) == y).sum().item()
            n += y.shape[0]
    net.train(is_training)  # 恢复net的train/eval状态
    return l_sum / n, acc_sum / n

def train(net, train_iter, valid_iter, num_epochs, lr, wd, device, lr_period,
          lr_decay):
    loss = nn.CrossEntropyLoss()
    optimizer = optim.SGD(net.fc.parameters(), lr=lr, momentum=0.9, weight_decay=wd)
    net = net.to(device)
    for epoch in range(num_epochs):
        train_l_sum, n, start = 0.0, 0, time.time()
        if epoch > 0 and epoch % lr_period == 0:  # 每lr_period个epoch，学习率衰减一次
            lr = lr * lr_decay
            for param_group in optimizer.param_groups:
                param_group['lr'] = lr
        for X, y in train_iter:
            X, y = X.to(device), y.to(device)
            optimizer.zero_grad()
            y_hat = net(X)
            l = loss(y_hat, y)
            l.backward()
            optimizer.step()
            train_l_sum += l.item() * y.shape[0]
            n += y.shape[0]
        time_s = "time %.2f sec" % (time.time() - start)
        if valid_iter is not None:
            valid_loss, valid_acc = evaluate_loss_acc(valid_iter, net, device)
            epoch_s = ("epoch %d, train loss %f, valid loss %f, valid acc %f, "
                       % (epoch + 1, train_l_sum / n, valid_loss, valid_acc))
        else:
            epoch_s = ("epoch %d, train loss %f, "
                       % (epoch + 1, train_l_sum / n))
        print(epoch_s + time_s + ', lr ' + str(lr))

6.在已划分的训练集、验证集上调参

用train数据集训练模型，通过观察在valid数据集上的损失与准确率来调整超参数

num_epochs, lr_period, lr_decay = 20, 10, 0.1
lr, wd = 0.03, 1e-4
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
net = get_net(device)
train(net, train_iter, valid_iter, num_epochs, lr, wd, device, lr_period, lr_decay)

'''
Downloading: "https://download.pytorch.org/models/resnet34-333f7ec4.pth" to /root/.cache/torch/checkpoints/resnet34-333f7ec4.pth
100%|██████████| 83.3M/83.3M [00:00<00:00, 207MB/s]
epoch 1, train loss 3.313305, time 69.19 sec, lr 0.03
epoch 2, train loss 1.466622, time 65.52 sec, lr 0.03
epoch 3, train loss 1.294536, time 64.86 sec, lr 0.03
epoch 4, train loss 1.240731, time 64.99 sec, lr 0.03
epoch 5, train loss 1.154168, time 64.83 sec, lr 0.03
epoch 6, train loss 1.159632, time 64.20 sec, lr 0.03
epoch 7, train loss 1.113965, time 63.91 sec, lr 0.03
epoch 8, train loss 1.099198, time 64.07 sec, lr 0.03
epoch 9, train loss 1.067925, time 63.68 sec, lr 0.03
epoch 10, train loss 1.071222, time 63.30 sec, lr 0.03
epoch 11, train loss 0.951122, time 62.85 sec, lr 0.003
epoch 12, train loss 0.913102, time 63.10 sec, lr 0.003
epoch 13, train loss 0.867716, time 66.08 sec, lr 0.003
epoch 14, train loss 0.892567, time 67.11 sec, lr 0.003
epoch 15, train loss 0.875075, time 64.89 sec, lr 0.003
epoch 16, train loss 0.899784, time 65.42 sec, lr 0.003
epoch 17, train loss 0.875754, time 64.65 sec, lr 0.003
epoch 18, train loss 0.875984, time 64.52 sec, lr 0.003
epoch 19, train loss 0.879278, time 63.99 sec, lr 0.003
epoch 20, train loss 0.883361, time 64.47 sec, lr 0.003
'''

7.在完整数据集上训练模型

# 使用上面的参数设置，在完整数据集上训练模型大致需要40-50分钟的时间
net = get_net(device)
train(net, train_valid_iter, None, num_epochs, lr, wd, device, lr_period, lr_decay)

8.对测试集分类并提交结果

用训练好的模型对测试数据进行预测。比赛要求对测试集中的每张图片，都要预测其属于各个类别的概率。

preds = []
for X, _ in test_iter:
    X = X.to(device)
    output = net(X)
    output = torch.softmax(output, dim=1)
    preds += output.tolist()
ids = sorted(os.listdir(os.path.join(new_data_dir, 'test/unknown')))
with open('submission.csv', 'w') as f:
    f.write('id,' + ','.join(train_valid_ds.classes) + '\n')
    for i, output in zip(ids, preds):
        f.write(i.split('.')[0] + ',' + ','.join(
            [str(num) for num in output]) + '\n')

课后习题

图像分类课后习题

狗品种识别课后习题