卷积神经网络

"""
手写字体的训练
"""
import os
import torch
import torch.nn as nn
from torch.autograd import Variable
import torch.utils.data as Data
import torchvision
import matplotlib.pyplot as plt

# 超参数
EPOCH = 1
BATCH_SIZE = 50
LR = 0.001
DOWNLOAD_MNIST = False

# 确认有没有mnist数据集
if not(os.path.exists('./mnist/')) or not os.listdir('./mnist/'):
    DOWNLOAD_MNIST = True

# 在mnist官网下载训练数据集
train_data = torchvision.datasets.MNIST(
    root='./mnist/', # 文件下载后的保存路径
    train=True, # True代表训练数据集，False代表测试数据集
    transform=torchvision.transforms.ToTensor(), # 将下载后的数据转换为tensor格式，并将数据归一化到0到1之间                                                    
    download=DOWNLOAD_MNIST,# 是否下载数据集
)

print(train_data.train_data.size())                 # (60000, 28, 28)
print(train_data.train_labels.size())               # (60000)

# 显示出第一张图片
plt.imshow(train_data.train_data[0].numpy(), cmap='gray')
plt.title('%i' % train_data.train_labels[0])
plt.show()

# 将以上数据集分为多批
train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)

# 测试数据集（还没有经过transform操作，其数据范围是0到255之间）
test_data = torchvision.datasets.MNIST(root='./mnist/', train=False)

# 将以下数据归一化到0到1之间，所以要除以255
test_x = Variable(torch.unsqueeze(test_data.test_data, dim=1), volatile=True).type(torch.FloatTensor)[:2000]/255.   # shape from (2000, 28, 28) to (2000, 1, 28, 28), value in range(0,1)
test_y = test_data.test_labels[:2000]


class CNN(nn.Module):
    def __init__(self):
        super(CNN, self).__init__()
        self.conv1 = nn.Sequential(         # 图片的形状为(1, 28, 28)
            nn.Conv2d(
                in_channels=1,              # 被卷积的通道数
                out_channels=16,            # 输出的通道数
                kernel_size=5,              # 卷积核的大小为（5x5）
                stride=1,                   # 卷积核的移动步数为1
                padding=2,                  # 图片的拓展圈数
            ),                              # 输出形状为(16, 28, 28)
            nn.ReLU(),                      # 激活函数
            nn.MaxPool2d(kernel_size=2),    # 最大池化后输出形状为(16, 14, 14)
        )
        self.conv2 = nn.Sequential(         # 输入形状为(16, 14, 14)
            nn.Conv2d(16, 32, 5, 1, 2),     # 输出形状为(32, 14, 14)
            nn.ReLU(),                      # 激活函数
            nn.MaxPool2d(2),                # 最大池化后输出形状为(32, 7, 7)
        )
        self.out = nn.Linear(32 * 7 * 7, 10)   # 全连接

    def forward(self, x):
        x = self.conv1(x)
        x = self.conv2(x)          # x的形状为(batch_size, 32, 7, 7)
        x = x.view(x.size(0), -1)  # 执行后x的形状为(batch_size, 32 * 7 * 7)
        output = self.out(x)
        return output


cnn = CNN()
print(cnn)  # 打印出网络结构

# 优化所有的网络参数
optimizer = torch.optim.Adam(cnn.parameters(), lr=LR)

#计算损失值
loss_func = nn.CrossEntropyLoss()

# 训练及测试
for epoch in range(EPOCH):
    for step, (x, y) in enumerate(train_loader):
        b_x = Variable(x)
        b_y = Variable(y)

        output = cnn(b_x)
        loss = loss_func(output, b_y)
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        if step % 50 == 0:
            test_output = cnn(test_x)
            pred_y = torch.max(test_output, 1)[1].data.squeeze()
            accuracy = sum(pred_y == test_y) / float(test_y.size(0))
            print('Epoch: ', epoch, '| train loss: %.4f' % loss.data[0], '| test accuracy: %.2f' % accuracy)

# 打印出前十个图片的预测效果
test_output, _ = cnn(test_x[:10])
pred_y = torch.max(test_output, 1)[1].data.numpy().squeeze()
print(pred_y, 'prediction number')
print(test_y[:10].numpy(), 'real number')