【人工智能概論】 構(gòu)建神經(jīng)網(wǎng)絡(luò)——以用InceptionNet解決MNIST任務(wù)為例

一. 整體思路

• 兩條原則，四個(gè)步驟。

1.1 兩條原則

1. 從宏觀到微觀
2. 把握數(shù)據(jù)形狀

1.2 四個(gè)步驟

1. 準(zhǔn)備數(shù)據(jù)
2. 構(gòu)建模型
3. 確定優(yōu)化策略
4. 完善訓(xùn)練與測(cè)試代碼

二. 舉例——用InceptionNet解決MNIST任務(wù)

2.1 模型簡(jiǎn)介

• InceptionNet的設(shè)計(jì)思路是通過(guò)增加網(wǎng)絡(luò)寬度來(lái)獲得更好的模型性能。
• 其核心在于基本單元Inception結(jié)構(gòu)塊，如下圖：
  
• 通過(guò)縱向堆疊Inception塊構(gòu)建完整網(wǎng)絡(luò)。

2.2 MNIST任務(wù)

• MNIST是入門級(jí)的機(jī)器學(xué)習(xí)任務(wù)；
• 它是一個(gè)手寫數(shù)字識(shí)別的數(shù)據(jù)集。

2.3 完整的程序

# 調(diào)包
import torch
import torch.nn as nn
import torch.nn.functional as F
from torchvision import transforms
from torchvision import datasets
from torch.utils.data import DataLoader
import torch.optim as optim


"""數(shù)據(jù)準(zhǔn)備"""
batch_size = 64

transform = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize((0.1307,),(0.3081,))
])

train_dataset = datasets.MNIST(root='./mnist/',train=True,download=True,transform=transform)
train_loader = DataLoader(train_dataset,shuffle=True,batch_size=batch_size)
test_dataset = datasets.MNIST(root='./mnist/',train=False,download=True,transform=transform)
test_loader = DataLoader(test_dataset,shuffle=False,batch_size=batch_size)


"""構(gòu)建模型"""
# 需要指定輸入的通道數(shù)

class Inceptiona(nn.Module): 
    def __init__(self,in_channels):
        super(Inceptiona,self).__init__()
        
        self.branch1_1 = nn.Conv2d(in_channels , 16 , kernel_size= 1)
        
        self.branch5_5_1 =nn.Conv2d(in_channels, 16, kernel_size= 1)
        self.branch5_5_2 =nn.Conv2d(16,24,kernel_size=5,padding=2)
        
        self.branch3_3_1 = nn.Conv2d(in_channels, 16,kernel_size=1)
        self.branch3_3_2 = nn.Conv2d(16,24,kernel_size=3,padding=1)
        self.branch3_3_3 = nn.Conv2d(24,24,kernel_size=3,padding=1)
        
        self.branch_pooling = nn.Conv2d(in_channels,24,kernel_size=1)
        
    def forward(self,x):
        x1 = self.branch1_1(x)
        
        x2 = self.branch5_5_1(x)
        x2 = self.branch5_5_2(x2)
        
        x3 = self.branch3_3_1(x)
        x3 = self.branch3_3_2(x3)
        x3 = self.branch3_3_3(x3)
        
        x4 = F.avg_pool2d(x,kernel_size=3,stride = 1, padding=1)
        x4 = self.branch_pooling(x4)
        
        outputs = [x1,x2,x3,x4]
        return torch.cat(outputs,dim=1)      
    
# 構(gòu)建完整的網(wǎng)絡(luò)
class Net(nn.Module):
    def __init__(self):
        super(Net,self).__init__()
        self.conv1 = nn.Conv2d(1,10,kernel_size=5)
        self.conv2 = nn.Conv2d(88,20,kernel_size=5)
        
        self.incep1 = Inceptiona(in_channels=10)
        self.incep2 = Inceptiona(in_channels=20)
        
        self.mp = nn.MaxPool2d(2)
        self.fc = nn.Linear(1408,10)
        
    def forward(self,x):
        batch_size = x.size(0)
        x = F.relu(self.mp(self.conv1(x)))
        x = self.incep1(x)
        x = F.relu(self.mp(self.conv2(x)))
        x = self.incep2(x)
        
        x = x.view(batch_size,-1)
        x = self.fc(x)
        return x


"""確定優(yōu)化策略"""
model = Net()
device = torch.device('cuda:0'if torch.cuda.is_available() else 'cpu')
model.to(device) # 指定設(shè)備

criterion = torch.nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(),lr=0.01,momentum=0.5)



"""完善訓(xùn)練與測(cè)試代碼"""
def train(epoch):
    running_loss = 0.0
    for batch_index, data in enumerate(train_loader,0):
        inputs, target = data
        # 把數(shù)據(jù)和模型送到同一個(gè)設(shè)備上
        inputs, target = inputs.to(device), target.to(device)
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs,target)
        loss.backward()
        optimizer.step()
        
        running_loss += loss.item() 
        # 用loss.item不會(huì)構(gòu)建計(jì)算圖，得到的不是張量，而是標(biāo)量
        if batch_index % 300 == 299:
            # 每三百組計(jì)算一次平均損失
            print('[%d,%5d] loss: %.3f' %(epoch+1,batch_index+1,running_loss/300))
            # 給出的是平均每一輪的損失
            running_loss = 0.0
            
def test():
    correct = 0
    total = 0
    with torch.no_grad(): 
        # 測(cè)試的環(huán)節(jié)不用求梯度
        for data in test_loader:
            images , labels = data
            images , labels = images.to(device),  labels.to(device)
            outputs = model(images)
            _, predicted = torch.max(outputs.data,dim=1)
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
    print('accuracy on test set: %d,%%'%(100*correct/total))
    return 100*correct/total  # 將測(cè)試的準(zhǔn)確率返回

# 執(zhí)行訓(xùn)練
if __name__=='__main__':
    score_best = 0
    for epoch in range(10):
        train(epoch)
        score = test()
        if score > score_best:
            score_best = score
            torch.save(model.state_dict(), "model.pth")

文章來(lái)源地址http://www.zghlxwxcb.cn/news/detail-419523.html

到了這里，關(guān)于【人工智能概論】 構(gòu)建神經(jīng)網(wǎng)絡(luò)——以用InceptionNet解決MNIST任務(wù)為例的文章就介紹完了。如果您還想了解更多內(nèi)容，請(qǐng)?jiān)谟疑辖撬阉鱐OY模板網(wǎng)以前的文章或繼續(xù)瀏覽下面的相關(guān)文章，希望大家以后多多支持TOY模板網(wǎng)！