Kaggle上的圖像分類（CIFAR-10)

比賽網址是https://www.kaggle.com/c/cifar-10

導入庫

# 本節的網絡需要較長的訓練時間
# 可以在Kaggle訪問：
# 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

獲取和組織數據集

比賽數據分爲訓練集和測試集。訓練集包含 50,000 圖片。測試集包含 300,000 圖片。兩個數據集中的圖像格式均爲PNG，高度和寬度均爲32像素，並具有三個顏色通道（RGB）。圖像涵蓋10個類別：飛機，汽車，鳥類，貓，鹿，狗，青蛙，馬，船和卡車。 爲了更容易上手，我們提供了上述數據集的小樣本。“ train_tiny.zip”包含 80 訓練樣本，而“ test_tiny.zip”包含100個測試樣本。它們的未壓縮文件夾名稱分別是“ train_tiny”和“ test_tiny”。

圖像增強

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)

data = [d[0].data.cpu().numpy() for d in trainset]
np.mean(data) 

# 圖像增強
transform_train = transforms.Compose([
    transforms.RandomCrop(32, padding=4),  #先四周填充0，再把圖像隨機裁剪成32*32
    transforms.RandomHorizontalFlip(),  #圖像一半的概率翻轉，一半的概率不翻轉
    transforms.ToTensor(),
    transforms.Normalize((0.4731, 0.4822, 0.4465), (0.2212, 0.1994, 0.2010)), #R,G,B每層的歸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)),
])           

導入數據集

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']

定義模型

ResNet-18網絡結構：ResNet全名Residual Network殘差網絡。Kaiming He 的《Deep Residual Learning for Image Recognition》獲得了CVPR最佳論文。他提出的深度殘差網絡在2015年可以說是洗刷了圖像方面的各大比賽，以絕對優勢取得了多個比賽的冠軍。而且它在保證網絡精度的前提下，將網絡的深度達到了152層，後來又進一步加到1000的深度。

class ResidualBlock(nn.Module):   # 我們定義網絡時一般是繼承的torch.nn.Module創建新的子類

    def __init__(self, inchannel, outchannel, stride=1):
        super(ResidualBlock, self).__init__()
        #torch.nn.Sequential是一個Sequential容器，模塊將按照構造函數中傳遞的順序添加到模塊中。
        self.left = nn.Sequential(
            nn.Conv2d(inchannel, outchannel, kernel_size=3, stride=stride, padding=1, bias=False), 
            # 添加第一個卷積層,調用了nn裏面的Conv2d（）
            nn.BatchNorm2d(outchannel), # 進行數據的歸一化處理
            nn.ReLU(inplace=True), # 修正線性單元，是一種人工神經網絡中常用的激活函數
            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)
            )
        #  便於之後的聯合,要判斷Y = self.left(X)的形狀是否與X相同

    def forward(self, x): # 將兩個模塊的特徵進行結合，並使用ReLU激活函數得到最終的特徵。
        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個3x3的卷積核代替7x7的卷積核，減少模型參數
            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)   #第一個ResidualBlock的步幅由make_layer的函數參數stride指定
        # ，後續的num_blocks-1個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)

訓練和測試

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

# 超參數設置
EPOCH = 20   #遍歷數據集次數
pre_epoch = 0  # 定義已經遍歷數據集的次數
LR = 0.1        #學習率

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

# 定義損失函數和優化方式
criterion = nn.CrossEntropyLoss()  #損失函數爲交叉熵，多用於多分類問題
optimizer = optim.SGD(net.parameters(), lr=LR, momentum=0.9, weight_decay=5e-4) 
#優化方式爲mini-batch momentum-SGD，並採用L2正則化（權重衰減）

# 訓練
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): 
            #用於將一個可遍歷的數據對象(如列表、元組或字符串)組合爲一個索引序列，同時列出數據和數據下標，
            #下標起始位置爲0，返回 enumerate(枚舉) 對象。
            
            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) #選出每一列中最大的值作爲預測結果
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
            # 每20個batch打印一次loss和準確率
            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)   

以上感謝伯禹學習平臺和DWhale組織