jupyter+pytorch+minist

深度学习小白，喜欢看通俗易懂的教程，推荐几个我参考学习的博文：
入门级Pytorch+MINIST数据集实现手写数字识别
使用Pytorch实现手写数字识别(Mnist数据集)
十分钟掌握Pytorch搭建神经网络的流程

import torch
import torchvision
from torch.utils.data import DataLoader

n_epochs = 3            #循环整个训练数据集的次数
batch_size_train = 64   #单次传递给程序用以训练的数据（样本）个数
batch_size_test = 1000
learning_rate = 0.01    #学习率
momentum = 0.5          #降低了损失函数到达最低点的速度(梯度下降的一种)
log_interval = 10
random_seed = 1
torch.manual_seed(random_seed)

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)#shuffle=True表示打乱测试集

examples = enumerate(test_loader)
batch_idx, (example_data, example_targets) = next(examples)
print(example_targets)     #图片实际对应的数字标签
print(example_data.shape)  #一批测试数据是一个形状张量

import matplotlib.pyplot as plt
fig = plt.figure()
for i in range(6):
  plt.subplot(2,3,i+1)
  plt.tight_layout()
  plt.imshow(example_data[i][0], cmap='gray', interpolation='none')
  plt.title("Ground Truth: {}".format(example_targets[i]))
  plt.xticks([])
  plt.yticks([])
plt.show()

'''
我们将使用两个2d卷积层，然后是两个全连接(或线性)层。
作为激活函数，我们将选择整流线性单元(简称ReLUs)，作为正则化的手段，我们将使用两个dropout层。
dropout层:防止神经网络过拟合的技术。
'''
#nn.Module是PyTorch提供的神经网络类，实现了网络各层的定义及前向计算与反向传播机制。
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim

class Net(nn.Module):
    #神经网络的各个层，网络结构
    def __init__(self):      
        #初始化父类nn.Module
        super(Net, self).__init__()   
        #输入1通道，输出10通道，卷积核5*5的的一个卷积层conv1
        self.conv1 = nn.Conv2d(1, 10, kernel_size=5)          
        self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
        self.conv2_drop = nn.Dropout2d()
        #2个全连接层
        self.fc1 = nn.Linear(320, 50) #定义Linear的输入和输出的形状
        self.fc2 = nn.Linear(50, 10)
    #前向传播方式
    def forward(self, x):
        #输入的x先经过定义的conv1，再经过激活函数F.relu()
        x = F.relu(F.max_pool2d(self.conv1(x), 2))
        x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
        x = x.view(-1, 320) #对数据形状变形，-1表示该位置根据后面的形状自动调整
        x = F.relu(self.fc1(x))
        x = F.dropout(x, training=self.training)
        x = self.fc2(x)
        return F.log_softmax(x)

network = Net()
optimizer = optim.SGD(network.parameters(), lr=learning_rate,
            momentum=momentum)            #随机梯度下降

def train(epoch):
  network.train()#训练模式
  for batch_idx, (data, target) in enumerate(train_loader):
    optimizer.zero_grad()                #上一轮梯度清零
    output = network(data)
    loss = F.nll_loss(output, target)    #损失函数
    loss.backward()                      #反向传播
    optimizer.step()                     #更新参数
    if batch_idx % log_interval == 0:
      print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
        epoch, batch_idx * len(data), len(train_loader.dataset),
        100. * batch_idx / len(train_loader), loss.item()))
      train_losses.append(loss.item())
      train_counter.append(
        (batch_idx*64) + ((epoch-1)*len(train_loader.dataset)))
      torch.save(network.state_dict(), './model.pth')      #保存模型参数
      torch.save(optimizer.state_dict(), './optimizer.pth')#保存优化器参数
          
train(1)

def test():
  network.eval()#验证模式 train()和eval()最主要就是影响了BN层和Dropout层
  test_loss = 0
  correct = 0
  #no_grad()禁用梯度
  with torch.no_grad():
    for data, target in test_loader:
      output = network(data)
      test_loss += F.nll_loss(output, target, size_average=False).item()
      pred = output.data.max(1, keepdim=True)[1]  #获取最大值的位置,[batch_size,1]
      correct += pred.eq(target.data.view_as(pred)).sum()
  test_loss /= len(test_loader.dataset)
  test_losses.append(test_loss)
  print('\nTest set: Avg. loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
    test_loss, correct, len(test_loader.dataset),
    100. * correct / len(test_loader.dataset)))
    
test()

examples = enumerate(test_loader)
batch_idx, (example_data, example_targets) = next(examples)
with torch.no_grad():
  output = network(example_data)
fig = plt.figure()
for i in range(6):
  plt.subplot(2,3,i+1)
  plt.tight_layout()
  plt.imshow(example_data[i][0], cmap='gray', interpolation='none')#显示样本图片
  plt.title("Prediction: {}".format(
    output.data.max(1, keepdim=True)[1][i].item()))
  plt.xticks([])#隐藏坐标轴
  plt.yticks([])
plt.show()