西瓜玲子5.27打卡日记

Datawhale 零基础入门CV赛事-Task3 字符识别模型

卷积神经网络（简称CNN) 是一类特殊的人工神经网络，是深度学习中重要的一个分支。CNN在很多领域都表现优异，精度和速度比传统计算学习算法高很多。特别是在计算机视觉领域，CNN是解决图像分类、图像检索、物体检测和语义分割的主流模型。

CNN每一层由众多的卷积核组成，每个卷积核对输入的像素进行卷积操作，得到下一次的输入。随着网络层的增加卷积核会逐渐扩大感受野，并缩减图像的尺寸。

CNN是一种层次模型，输入的是原始的像素数据。CNN通过卷积（convolution）、池化（pooling）、非线性激活函数（non-linear activation function）和全连接层（fully connected layer）构成。

本次实验中构建模型代码如下：

import torch
torch.manual_seed(0)
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.nn.functional as F
import torch.optim as optim
from torch.autograd import Variable
from torch.utils.data.dataset import Dataset

# 定义模型
class SVHN_Model1(nn.Module):
    def __init__(self):
        super(SVHN_Model1, self).__init__()
        # CNN提取特征模块
        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),
        )
        # 
        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)
        self.fc6 = nn.Linear(32*3*7, 11)
    
    def forward(self, img):        
        feat = self.cnn(img)
        feat = feat.view(feat.shape[0], -1)
        c1 = self.fc1(feat)
        c2 = self.fc2(feat)
        c3 = self.fc3(feat)
        c4 = self.fc4(feat)
        c5 = self.fc5(feat)
        c6 = self.fc6(feat)
        return c1, c2, c3, c4, c5, c6
    
model = SVHN_Model1()

构建后输出模型如下：

SVHN_Model1(
  (cnn): Sequential(
    (0): Conv2d(3, 16, kernel_size=(3, 3), stride=(2, 2))
    (1): ReLU()
    (2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (3): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2))
    (4): ReLU()
    (5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  )
  (fc1): Linear(in_features=672, out_features=11, bias=True)
  (fc2): Linear(in_features=672, out_features=11, bias=True)
  (fc3): Linear(in_features=672, out_features=11, bias=True)
  (fc4): Linear(in_features=672, out_features=11, bias=True)
  (fc5): Linear(in_features=672, out_features=11, bias=True)
  (fc6): Linear(in_features=672, out_features=11, bias=True)
)

训练代码如下：

# 损失函数
criterion = nn.CrossEntropyLoss()
# 优化器
optimizer = torch.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, c5 = 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]) + \
                criterion(c5, data[1][:, 5])
        loss /= 6
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        
        loss_plot.append(loss.item())
        c0_plot.append((c0.argmax(1) == data[1][:, 0]).sum().item()*1.0 / c0.shape[0])
        
    print(epoch)

运行代码后存在bug，将loss部分定义改为（将数据类型改为long）：

        loss = criterion(c0, data[1][:, 0].long()) + \
                criterion(c1, data[1][:, 1].long()) + \
                criterion(c2, data[1][:, 2].long()) + \
                criterion(c3, data[1][:, 3].long()) + \
                criterion(c4, data[1][:, 4].long()) + \
                criterion(c5, data[1][:, 5].long())

将

c0_plot.append((c0.argmax(1) == data[1][:, 0]).sum().item()*1.0 / c0.shape[0])

改为:

c0_plot.append((c0.argmax(1) == data[1][:, 0].long()).sum().item()*1.0 / c0.shape[0])

同时Dataloader的代码也需要修改，将以下代码中的5改为6。

lbl = list(lbl)  + (5 - len(lbl)) * [10]

修改完毕后，代码可运行开始训练。