天池-零基础入门CV-街景字符编码识别-字符识别模型

CNN及典型网络介绍

• 采用PyTorch构建CNN模型
  在Pytorch中构建CNN模型非常简单，只需要定义好模型的参数和正向传播即可，PyTorch会根据正向传播自动计算反向传播。以下以构造一个简单的CNN为例，然后进行训练。这个CNN模型包括两个卷积层，最后并联6个全连接层进行分类。

import torch
torch.manual_seed(0) #设置初始化
torch.backends.cudnn.deterministic = False
torch.backends.cudnn.benchmark = True

import torchvision.models as models
import torchvision.transforms as transforms
import torchvision.datasets as datasets
import torch.nn as nn
import torch.optim as optim
from torch.autograd import Variable
from torch.utils.data.dataset import Dataset

#定义模型
class SVNH_Model1(nn.Module):
    def __init__(self):
        super(SVNH_Model1, self).__init__()
        #CNN特征提取模块
        self.cnn = nn.Sequential(
            nn.Conv2d(3, 16, kernel_size=(3, 3), stride=(2, 2)),
            nn.ReLU(),  
            nn.MaxPool2d(2),
            nn.Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2)),
            nn.ReLU(), 
            nn.MaxPool2d(2),
        )
        #6个并行的全连接层
        self.fc1 = nn.Linear(32*3*7, 11)
        self.fc2 = nn.Linear(32*3*7, 11)
        self.fc3 = nn.Linear(32*3*7, 11)
        self.fc4 = nn.Linear(32*3*7, 11)
        self.fc5 = nn.Linear(32*3*7, 11)
        
        
    def forward(self, img):        
        feat = self.cnn(img)
        print(feat.shape[0])
        feat = feat.view(feat.shape[0], -1) ### 将CNN的输出
        c1 = self.fc1(feat)
        c2 = self.fc2(feat)
        c3 = self.fc3(feat)
        c4 = self.fc4(feat)
        c5 = self.fc5(feat)
        return c1, c2, c3, c4, c5
    
model = SVNH_Model1()

# 训练代码
#损失函数
criterion = nn.CrossEntropyLoss()
#优化器
optimizer = optim.Adam(model.parameters(), 0.005)

loss_plot, c0_plot = [], []
# 迭代10个epoch
for epoch in range(10):
    for data in train_loader:
        c0, c1, c2, c3, c4 = model(data[0])
        loss = criterion(c0, data[1][:, 0]) + \
                criterion(c1, data[1][:, 1]) + \
                criterion(c2, data[1][:, 2]) + \
                criterion(c3, data[1][:, 3]) + \
                criterion(c4, data[1][:, 4]) 
        loss /= 5
        optimizer.zero_grad() #梯度置零
        loss.backward()
        optimizer.step() #更新模型
        
        loss_plot.append(loss.item())
        print(epoch)

### 绘制loss曲线
import numpy as np
import matplotlib.pyplot as plt
import numpy as np
import matplotlib.pyplot as plt
x = np.linspace(0, epoch, 7500)
plt.plot(x,loss_plot,label="loss",color="red",linewidth=2)
plt.xlabel("epoch")
plt.ylabel("loss")
plt.title("loss")
plt.ylim(0,3)
plt.show()

loss 曲线绘制如下，可看到训练基本收敛。

后续拟采用更加复杂的模型来做。