MNIST softmax

导包

import torch
import torch.nn as nn
import torchvision
import torchvision.transforms as transforms
from torchvision.datasets import MNIST
 
 #Hyper-parameters 
input_size = 784 #输入层神经元大小
num_classes = 10 #种类类别，数字0-9
num_epochs = 5#迭代次数
batch_size_train = 100 #每次训练取得样本数
batch_size_test=1000
learning_rate = 0.01 #学习率

准备数据

# MNIST dataset (images and labels)
#训练集
train_dataset = torchvision.datasets.MNIST(root='data/', 
                                           train=True, 
                                           transform=transforms.ToTensor(),
                                           download=True)
#测试集
test_dataset = torchvision.datasets.MNIST(root='data/', 
                                          train=False, 
                                          transform=transforms.ToTensor())
 
# Data loader (input pipeline)
train_loader = torch.utils.data.DataLoader(
                        torchvision.datasets.MNIST('./data/',train=True,download=True,
                        transform=torchvision.transforms.Compose([
                        torchvision.transforms.ToTensor(),
                        torchvision.transforms.Normalize(
                                 (0.1307,), (0.3081,))
                             ])),
  batch_size=batch_size_train, shuffle=True)
 
test_loader = torch.utils.data.DataLoader(
                        torchvision.datasets.MNIST('./data/',train=False,download=True,
                        transform=torchvision.transforms.Compose([
                        torchvision.transforms.ToTensor(),
                        torchvision.transforms.Normalize(
                                     (0.1307,), (0.3081,))
                             ])),
  batch_size=batch_size_test, shuffle=True)

网络

#Logistic regression model
model = nn.Linear(input_size, num_classes)
#Loss and optimizer 损失函数和优化器
# nn.CrossEntropyLoss() computes softmax internally
criterion = nn.CrossEntropyLoss()  #采用多分类交叉熵损失函数
optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate,momentum=0.5)  

训练

# Train the model 训练模型
total_step = len(train_loader)
total_losses=[]
step=[i*10000 for i in range(1,num_epochs*600+1)]
for epoch in range(num_epochs):
    for i, (images, labels) in enumerate(train_loader):
        # Reshape images to (batch_size, input_size)  
        images = images.reshape(-1, 28*28)
        
        # Forward pass 前向传播
        outputs = model(images)  #100x10
       
        loss = criterion(outputs, labels)
        total_losses.append(loss.item())
        
        # Backward and optimize 反向传播
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        
        if (i+1) % 100 == 0:
            print('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}' 
                   .format(epoch+1,num_epochs,i+1,total_step,loss.item()))

#可视化
import matplotlib.pyplot as plt
fig = plt.figure()
plt.plot(step,total_losses,color='blue')
plt.legend(['Train Loss'],loc='upper right')
plt.xlabel("number of training examples seen")
plt.ylabel("negative log likelihood loss")
plt.show()

测试模型

#Test the model
# In test phase, we don't need to compute gradients (for memory efficiency)
with torch.no_grad():
    correct = 0
    total = 0
    for images, labels in test_loader:
        images = images.reshape(-1, 28*28)
        outputs = model(images)
        _, predicted = torch.max(outputs.data, 1)
        total += labels.size(0)
        correct += (predicted == labels).sum()
 
    print('Accuracy of the model on the 10000 test images: {} %'.format(100 * correct / total))
 #Save the model checkpoint
torch.save(model.state_dict(), 'model.ckpt')

导入模型的时候，需要把网络模型写出来，再导入，直接model=torch.load('model.ckpt')会报错

model = nn.Linear(784, 10)
model.load_state_dict(torch.load('model.ckpt'))#再加载网络的参数
print("load success")