刘二大人第九讲

# import numpy as np
# y = np.array([1,0,0])
# z = np.array([0.2,0.1,-0.1])
# y_pred = np.exp(z) / np.exp(z).sum()
# loss = (-y * np.log(y_pred)).sum()
# print(loss)


# import torch
# criterion = torch.nn.CrossEntropyLoss()
# Y = torch.LongTensor([2,0,1])
# Y_pred1 = torch.Tensor([[0.1,0.2,0.9],
#                         [1.1,0.1,0.2],
#                         [0.2,2.1,0.1]])
# Y_pred2 = torch.Tensor([[0.9,0.2,0.4],
#                         [0.2,0.9,2.0],
#                         [0.2,0.7,0.9]])
#
# l1 = criterion(Y_pred1,Y)
# l2 = criterion(Y_pred2,Y)
# print("Batch Loss1 = ",l1.data,"Batch Loss2 = ",l2.data)



import torch


#下面三个和数据集有关的包
import torchvision
from torchvision import transforms
from torchvision import datasets
from torch.utils.data import DataLoader



import torch.nn.functional as F #激活函数使用ReLU
import torch.optim as optim #使用优化器

batch_size = 64


#转换为图像张量
transforms = transforms.Compose([
    transforms.ToTensor(),  #convert the PIL Image to Tensor,把单通道变成多通道
    transforms.Normalize((0.1307, ), (0.3081, )) #归一化，切换到0-1分布，前面是均值，后面是方差
])



train_dataset = torchvision.datasets.MNIST(
    root="../data/MNIST",
    train=True, #只使用训练集
    #将数据转化为torch使用的张量，取值范围为[0,1]
    transform=torchvision.transforms.ToTensor(),#参数transfrom用于指定数据集的变换，
#transform=transforms.ToTensor()表示将数据中的像素值转换到0~1之间，并且将图像数据从形状为[H,W,C]转换成形状为[C,H,W]
    download=False
)

# train_dataset = datasets.MNIST(root='../data/MNIST/',
#                                train=True,
#                                transform=transforms)
train_loader = DataLoader(train_dataset,
                          shuffle=True,
                          batch_size=batch_size)
test_dataset = datasets.MNIST(root='../data/MNIST/',
                              train=False,
                              download=False,
                              transform=transforms
                              )
test_loader = DataLoader(test_dataset,
                         shuffle=False,
                         batch_size=batch_size)

#N个样本，每个样本是1维28×28的图像(N,1,28,28)

#输入层
# x = x.view(-1,784) #使用view函数，把张量变成二阶张量，也就是矩阵，列数是784.28×28=784
#
# #线性层
# self.l1 = torch.nn.Linear(784,512)
#
# #ReLU层
# x = F.relu(self.l1(x)),用ReLU对每一层进行激活

class Net(torch.nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.l1 = torch.nn.Linear(784,512)
        self.l2 = torch.nn.Linear(512,256)
        self.l3 = torch.nn.Linear(256,128)
        self.l4 = torch.nn.Linear(128,64)
        self.l5 = torch.nn.Linear(64,10)

    def forward(self,x):
        x = x.view(-1,784)
        x = F.relu(self.l1(x))
        x = F.relu(self.l2(x))
        x = F.relu(self.l3(x))
        x = F.relu(self.l4(x))
        return self.l5(x)
model = Net()


#3
criterion = torch.nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(),lr=0.01,momentum=0.5) #因为模型比较大，所以用带冲量的，momentum=0.5，来优化训练过程

#4

#把单独的一轮封装到函数里
def train(epoch):
    running_loss = 0.0
    for batch_idx,data in enumerate(train_loader,0):
        inputs,target = data #inputs是x，target是y
        optimizer.zero_grad()

        #forward + backward + update
        outputs = model(inputs)
        loss = criterion(outputs,target)
        loss.backward()
        optimizer.step()

        running_loss += loss.item()
        if batch_idx % 300 == 299: #每训练300次输出
            print('[%d, %5d loss:%.3f' % (epoch + 1,batch_idx + 1, running_loss/300))
            running_loss = 0.0


def test():
    correct = 0
    total = 0
    with torch.no_grad(): #不用计算梯度
        for data in test_loader:
            images,labels=data
            outputs = model(images)
            _,predicted = torch.max(outputs.data,dim=1) #求每一行最大值，也就是分类，返回每一行的最大值，以及最大值的下标
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
    print('Accuarcy on test set:%d %%' % (100*correct/total))

#训练
if __name__=='__main__':
    for epoch in range(10):
        train(epoch)
        test()