pytorch 用Lenet5實現MNIST手寫數字識別，迭代100次，正確率99.32%

1.訓練模型

import torch
import torchvision
from torchvision import datasets,transforms
import matplotlib.pyplot as plt
import numpy as np
import cv2

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

transform = transforms.Compose([
        transforms.ToTensor(), # 轉爲Tensor
        transforms.Normalize((0.5,), (0.5,)), # 歸一化
                             ])

train_dataset = torchvision.datasets.MNIST(root='./mnist', train=True, transform=transform, download=True)
test_dataset = torchvision.datasets.MNIST(root='./mnist', train=False, transform=transform, download=True)

batch_size = 4
trainloader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=False, num_workers=4)
testloader = torch.utils.data.DataLoader(test_dataset, batch_size=batch_size, shuffle=False, num_workers=4)

import torch.nn as nn
import torch.nn.functional as F

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 6, 5, padding=2)
        self.conv2 = nn.Conv2d(6, 16, 5)  
        self.fc1   = nn.Linear(16*5*5, 120)  
        self.fc2   = nn.Linear(120, 84)
        self.fc3   = nn.Linear(84, 10)

    def forward(self, x): 
        x = F.max_pool2d(F.relu(self.conv1(x)), (2,2)) 
        x = F.max_pool2d(F.relu(self.conv2(x)), 2) 
        x = x.view(x.size()[0], -1)   #展開成一維的
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)        
        return x

net = Net()
net.to(device)
print(net)

from torch import optim
criterion = nn.CrossEntropyLoss() # 交叉熵損失函數
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9) # 優化器

import time

start_time = time.time()

for epoch in range(100):
    running_loss = 0.0 #初始化loss
    for i, (inputs, labels) in enumerate(trainloader, 0):

        # 輸入數據

        inputs =  inputs.to(device)
        labels =  labels.to(device)

        # 梯度清零
        optimizer.zero_grad()

        # forward + backward 
        outputs = net(inputs)
        loss = criterion(outputs, labels)
        loss.backward()   

        # 更新參數 
        optimizer.step()

        # 打印log信息
        # loss 是一個scalar,需要使用loss.item()來獲取數值，不能使用loss[0]
        running_loss += loss.item()
        if i % 2000 == 1999: # 每2000個batch打印一下訓練狀態
            print('[%d, %5d] loss: %.3f' % (epoch+1, i+1, running_loss / 2000))
            running_loss = 0.0

stop_time = time.time()

print('Finished Training 耗時： ', (stop_time - start_time), '秒')

2.保存模型

PATH = './mnist_net_100.pth'
torch.save(net.state_dict(), PATH)

3.讀取並加載保存的模型

pretrained_net = torch.load(PATH)

net2 = Net()

net2.load_state_dict(pretrained_net)

4.在測試集上進行測試

#整個測試集上預測
correct = 0
total = 0

with torch.no_grad():
    for (images,labels) in testloader:
        images = images.to(device)
        labels = labels.to(device)
        outputs = net(images)
        _, predicted = torch.max(outputs,1)
        
        total += labels.size(0)
        correct += (predicted == labels).sum()

print('10000張測試集合中的準確率爲：', (correct.cpu().numpy()/total * 100))
print(correct)

5.抽取數據進行識別和顯示

classes = ('0', '1', '2', '3', '4', '5', '6', '7', '8', '9')

#**測試圖像的實際labels**
dataiter = iter(testloader) #把測試數據放在迭代器iter
images, labels = dataiter.next() # 一個batch返回4張圖片，依次獲取下一個數據
images = images.to(device)
labels = labels.to(device)
print('實際的label: ', ' '.join( '%08s'%classes[labels[j]] for j in range(4)))

print(images/2+0.5)
img = np.empty((28,28*4), dtype=np.float32)
img[:,0:28] = images[0].numpy()
img[:,28:56] = images[1].numpy()
img[:,56:84] = images[2].numpy()
img[:,84:112] = images[3].numpy()

img2 = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)
print(img2.shape)
plt.imshow(img2)
plt.show()