pytorch VGG11识别cifar10数据集(训练+预测单张输入图片操作)

首先这是VGG的结构图，VGG11则是红色框里的结构，共分五个block，如红框中的VGG11第一个block就是一个conv3-64卷积层：

一，写VGG代码时，首先定义一个 vgg_block(n,in,out)方法，用来构建VGG中每个block中的卷积核和池化层：

n是这个block中卷积层的数目，in是输入的通道数，out是输出的通道数

有了block以后，我们还需要一个方法把形成的block叠在一起，我们定义这个方法叫vgg_stack：

def vgg_stack(num_convs, channels): # vgg_net = vgg_stack((1, 1, 2, 2, 2), ((3, 64), (64, 128), (128, 256), (256, 512), (512, 512)))

net = []
for n, c in zip(num_convs, channels):
 in_c = c[0]
 out_c = c[1]
 net.append(vgg_block(n, in_c, out_c))
return nn.Sequential(*net)

右边的注释

vgg_net = vgg_stack((1, 1, 2, 2, 2), ((3, 64), (64, 128), (128, 256), (256, 512), (512, 512)))

里，(1, 1, 2, 2, 2)表示五个block里，各自的卷积层数目，((3, 64), (64, 128), (128, 256), (256, 512), (512, 512))表示每个block中的卷积层的类型，如(3,64)表示这个卷积层输入通道数是3，输出通道数是64。vgg_stack方法返回的就是完整的vgg11模型了。

接着定义一个vgg类，包含vgg_stack方法：

#vgg类
class vgg(nn.Module):
def __init__(self):
 super(vgg, self).__init__()
 self.feature = vgg_net
 self.fc = nn.Sequential(
  nn.Linear(512, 100),
  nn.ReLU(True),
  nn.Linear(100, 10)
 )

def forward(self, x):
 x = self.feature(x)
 x = x.view(x.shape[0], -1)
 x = self.fc(x)
 return x

最后：

net = vgg() #就能获取到vgg网络

那么构建vgg网络完整的pytorch代码是：

def vgg_block(num_convs, in_channels, out_channels):
net = [nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1), nn.ReLU(True)]

for i in range(num_convs - 1): # 定义后面的许多层
 net.append(nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1))
 net.append(nn.ReLU(True))

net.append(nn.MaxPool2d(2, 2)) # 定义池化层
return nn.Sequential(*net)

# 下面我们定义一个函数对这个 vgg block 进行堆叠
def vgg_stack(num_convs, channels): # vgg_net = vgg_stack((1, 1, 2, 2, 2), ((3, 64), (64, 128), (128, 256), (256, 512), (512, 512)))
net = []
for n, c in zip(num_convs, channels):
 in_c = c[0]
 out_c = c[1]
 net.append(vgg_block(n, in_c, out_c))
return nn.Sequential(*net)

#确定vgg的类型，是vgg11 还是vgg16还是vgg19
vgg_net = vgg_stack((1, 1, 2, 2, 2), ((3, 64), (64, 128), (128, 256), (256, 512), (512, 512)))
#vgg类
class vgg(nn.Module):
def __init__(self):
 super(vgg, self).__init__()
 self.feature = vgg_net
 self.fc = nn.Sequential(
  nn.Linear(512, 100),
  nn.ReLU(True),
  nn.Linear(100, 10)
 )
def forward(self, x):
 x = self.feature(x)
 x = x.view(x.shape[0], -1)
 x = self.fc(x)
 return x

#获取vgg网络
net = vgg()

基于VGG11的cifar10训练代码：

import sys
import numpy as np
import torch
from torch import nn
from torch.autograd import Variable
from torchvision.datasets import CIFAR10
import torchvision.transforms as transforms

def vgg_block(num_convs, in_channels, out_channels):
net = [nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1), nn.ReLU(True)]

for i in range(num_convs - 1): # 定义后面的许多层
 net.append(nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1))
 net.append(nn.ReLU(True))

net.append(nn.MaxPool2d(2, 2)) # 定义池化层
return nn.Sequential(*net)

# 下面我们定义一个函数对这个 vgg block 进行堆叠
def vgg_stack(num_convs, channels): # vgg_net = vgg_stack((1, 1, 2, 2, 2), ((3, 64), (64, 128), (128, 256), (256, 512), (512, 512)))
net = []
for n, c in zip(num_convs, channels):
 in_c = c[0]
 out_c = c[1]
 net.append(vgg_block(n, in_c, out_c))
return nn.Sequential(*net)

#vgg类
class vgg(nn.Module):
def __init__(self):
 super(vgg, self).__init__()
 self.feature = vgg_net
 self.fc = nn.Sequential(
  nn.Linear(512, 100),
  nn.ReLU(True),
  nn.Linear(100, 10)
 )
def forward(self, x):
 x = self.feature(x)
 x = x.view(x.shape[0], -1)
 x = self.fc(x)
 return x

# 然后我们可以训练我们的模型看看在 cifar10 上的效果
def data_tf(x):
x = np.array(x, dtype='float32') / 255
x = (x - 0.5) / 0.5
x = x.transpose((2, 0, 1)) ## 将 channel 放到第一维，只是 pytorch 要求的输入方式
x = torch.from_numpy(x)
return x

transform = transforms.Compose([transforms.ToTensor(),
        transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5)),
        ])
def get_acc(output, label):
total = output.shape[0]
_, pred_label = output.max(1)
num_correct = (pred_label == label).sum().item()
return num_correct / total

def train(net, train_data, valid_data, num_epochs, optimizer, criterion):
if torch.cuda.is_available():
 net = net.cuda()
for epoch in range(num_epochs):
 train_loss = 0
 train_acc = 0
 net = net.train()
 for im, label in train_data:
  if torch.cuda.is_available():
   im = Variable(im.cuda())
   label = Variable(label.cuda())
  else:
   im = Variable(im)
   label = Variable(label)
  # forward
  output = net(im)
  loss = criterion(output, label)
  # forward
  optimizer.zero_grad()
  loss.backward()
  optimizer.step()

train_loss += loss.item()
  train_acc += get_acc(output, label)

if valid_data is not None:
  valid_loss = 0
  valid_acc = 0
  net = net.eval()
  for im, label in valid_data:
   if torch.cuda.is_available():
    with torch.no_grad():
     im = Variable(im.cuda())
     label = Variable(label.cuda())
   else:
    with torch.no_grad():
     im = Variable(im)
     label = Variable(label)
   output = net(im)
   loss = criterion(output, label)
   valid_loss += loss.item()
   valid_acc += get_acc(output, label)
  epoch_str = (
    "Epoch ％d. Train Loss: ％f, Train Acc: ％f, Valid Loss: ％f, Valid Acc: ％f, "
    ％ (epoch, train_loss / len(train_data),
     train_acc / len(train_data), valid_loss / len(valid_data),
     valid_acc / len(valid_data)))
 else:
  epoch_str = ("Epoch ％d. Train Loss: ％f, Train Acc: ％f, " ％
      (epoch, train_loss / len(train_data),
      train_acc / len(train_data)))

# prev_time = cur_time
 print(epoch_str)

if __name__ == '__main__':
# 作为实例，我们定义一个稍微简单一点的 vgg11 结构，其中有 8 个卷积层
vgg_net = vgg_stack((1, 1, 2, 2, 2), ((3, 64), (64, 128), (128, 256), (256, 512), (512, 512)))
print(vgg_net)

train_set = CIFAR10('./data', train=True, transform=transform, download=True)
train_data = torch.utils.data.DataLoader(train_set, batch_size=64, shuffle=True)
test_set = CIFAR10('./data', train=False, transform=transform, download=True)
test_data = torch.utils.data.DataLoader(test_set, batch_size=128, shuffle=False)

net = vgg()
optimizer = torch.optim.SGD(net.parameters(), lr=1e-1)
criterion = nn.CrossEntropyLoss() #损失函数为交叉熵

train(net, train_data, test_data, 50, optimizer, criterion)
torch.save(net, 'vgg_model.pth')

结束后，会出现一个模型文件vgg_model.pth

二，然后网上找张图片，把图片缩成32x32，放到预测代码中，即可有预测结果出现，预测代码如下：

import torch
import cv2
import torch.nn.functional as F
from vgg2 import vgg ##重要，虽然显示灰色(即在次代码中没用到)，但若没有引入这个模型代码，加载模型时会找不到模型
from torch.autograd import Variable
from torchvision import datasets, transforms
import numpy as np

classes = ('plane', 'car', 'bird', 'cat',
  'deer', 'dog', 'frog', 'horse', 'ship', 'truck')
if __name__ == '__main__':
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = torch.load('vgg_model.pth') # 加载模型
model = model.to(device)
model.eval() # 把模型转为test模式

img = cv2.imread("horse.jpg") # 读取要预测的图片
trans = transforms.Compose(
 [
  transforms.ToTensor(),
  transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5))
 ])

img = trans(img)
img = img.to(device)
img = img.unsqueeze(0) # 图片扩展多一维,因为输入到保存的模型中是4维的[batch_size,通道,长，宽]，而普通图片只有三维，[通道,长，宽]
# 扩展后，为[1，1，28，28]
output = model(img)
prob = F.softmax(output,dim=1) #prob是10个分类的概率
print(prob)
value, predicted = torch.max(output.data, 1)
print(predicted.item())
print(value)
pred_class = classes[predicted.item()]
print(pred_class)

# prob = F.softmax(output, dim=1)
# prob = Variable(prob)
# prob = prob.cpu().numpy() # 用GPU的数据训练的模型保存的参数都是gpu形式的，要显示则先要转回cpu，再转回numpy模式
# print(prob) # prob是10个分类的概率
# pred = np.argmax(prob) # 选出概率最大的一个
# # print(pred)
# # print(pred.item())
# pred_class = classes[pred]
# print(pred_class)

缩成32x32的图片：

运行结果：

来源：https://blog.csdn.net/u014453898/article/details/91380837