刘二大人《PyTorch深度学习实践》完结合集——第9课：多分类问题

import torch
from torchvision import transforms                #针对图像进行处理
from  torchvision import datasets
from torch.utils.data import  DataLoader          #loader picture
import torch.nn.functional as F                   #using function rule
import torch.optim as optim
import matplotlib.pyplot as plt
# 糖尿病预测研判..................................................................................................#
#loader dataset and transform dataset as Tensore
batch_size = 64

transform = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.1307, ),(0.3081, ))])
#mnist库的训练集当中有60000张图片
train_dataset = datasets.MNIST(root='../dataset/mnist/',train=True,download=True,transform=transform)
train_loader =  DataLoader(train_dataset,shuffle=True,batch_size=batch_size )
#mnist库的训练集当中有10000张图片
test_dataset = datasets.MNIST(root='../dataset/mnist/',train=False,download=True,transform=transform)
test_loader = DataLoader(test_dataset,shuffle=False,batch_size=batch_size )


#Design model and use class.................. inherit from nn.Moduel........................................................#
class  Model(torch.nn.Module):
    def __init__(self):
        super( Model, 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)          # 图像的类型为 tensor 张量 （1，28，28）进行把图像平展化，变成一维张量
        x = F.relu(self.l1(x))      #input：784    output:512
        x = F.relu(self.l2(x))      #input：512    output:256
        x = F.relu(self.l3(x))      #input：256    output:128
        x = F.relu(self.l4(x))      #input：128    output:64
        x = self.l5(x)              #input：64    output:10
        return x

model =  Model()#实例化

#Construct loss functions and optimizer.................Use Torch API...................................................#
criterion  =  torch.nn.CrossEntropyLoss()     #老师使用的是torch.nn.BCELoss(size_average=False)但是我使用这个损失太大了
optimizer =   optim.SGD(model.parameters(),lr=0.01,momentum=0.5) #lr为学习率，因为0.01太小了，我改成了0.1

#plot
cycle = 10
x_axis = []
for i in range(cycle*3):
    x_axis.append((i+1))
y_axis = []

#defining one training..........................................................................................................#
def train(epoch):
    running_loss =0.0
    for batch_idx ,data in enumerate(train_loader,0):   #enumerate函数将train_loader中的元组(mini_batch(四维张量)+target(一维张量))列表化，index从0开始
        inputs ,target =data
        optimizer.zero_grad()

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

        running_loss +=loss.item()     #不用item，pytorch会构建计算图
        if batch_idx % 300 ==299:
            print('[%d,%5d] loss:%.3f' %(epoch+1,batch_idx+1,running_loss/300))
            y_axis.append(running_loss/300)
            running_loss = 0.0
#defining one testing..........................................................................................................#
def test():
    correct = 0
    total = 0
    with torch.no_grad():
        for data in test_loader:
            images,labels = data
            outputs = model(images)
            max,predicted = torch.max(outputs.data,dim=1)
            total = total+labels.size(0)
            correct =correct+(predicted == labels).sum().item()
    print('Accuracy on test set:%d %%' %(100*correct/total))

if __name__ == '__main__':
    for epoch in range(cycle):
        train(epoch)
        test()

#drawing.....................................................................................................#
plt.figure(figsize=(7, 7), dpi=80)  # 创建画布
plt.plot(x_axis,y_axis, color='r', linestyle='-', label='loss value&No.i mini_batch', )  # 绘制折线图，点划线
plt.xlabel('mini_batch')                                                       #设置图x轴标签
plt.ylabel('loss')                                                        #设置图y轴标签
plt.show()                                                                #显示图

D:\Anaconda\envs\study\python.exe "D:\python pycharm learning\刘二大人课程\P\P9.py" 
[1,  300] loss:2.141
[1,  600] loss:0.737
[1,  900] loss:0.415
Accuracy on test set:89 %
[2,  300] loss:0.310
[2,  600] loss:0.272
[2,  900] loss:0.232
Accuracy on test set:94 %
[3,  300] loss:0.195
[3,  600] loss:0.167
[3,  900] loss:0.148
Accuracy on test set:95 %
[4,  300] loss:0.128
[4,  600] loss:0.125
[4,  900] loss:0.115
Accuracy on test set:96 %
[5,  300] loss:0.094
[5,  600] loss:0.096
[5,  900] loss:0.096
Accuracy on test set:97 %
[6,  300] loss:0.074
[6,  600] loss:0.078
[6,  900] loss:0.074
Accuracy on test set:97 %
[7,  300] loss:0.059
[7,  600] loss:0.061
[7,  900] loss:0.062
Accuracy on test set:96 %
[8,  300] loss:0.049
[8,  600] loss:0.048
[8,  900] loss:0.053
Accuracy on test set:97 %
[9,  300] loss:0.040
[9,  600] loss:0.042
[9,  900] loss:0.040
Accuracy on test set:97 %
[10,  300] loss:0.030
[10,  600] loss:0.034
[10,  900] loss:0.036
Accuracy on test set:97 %

补充问题：我想在CUDA上面跑，所以在代码中修改：

data = data.to(device)
inputs, target = data
#inputs = inputs.to(device)
#target = target.to(device)  发现报错，根据ChatGTP提供的原因：