def train_model(model,train_loader):
    optimizer = torch.optim.Adam(model.parameters())
    loss_func = nn.CrossEntropyLoss()
    EPOCHS = 5
    for epoch in range(EPOCHS):
        correct = 0
        for batch_idx,(X_batch,y_batch) in enumerate(train_loader):
            optimizer.zero_grad()
            #這里是只取訓(xùn)練數(shù)據(jù)的意思嗎，X_batch和y_batch是怎么分開(kāi)的?
            #答：X_batch和y_batch是一一對(duì)應(yīng)的，只不過(guò)順序打亂了，參考torch.utils.data.ipynb
            output = model(X_batch.float())   #X_batch.float()是什么意思
            loss = loss_func(output,y_batch)
            loss.backward()
            optimizer.step()
            
            # Total correct predictions
            #第一個(gè)1代表取每行的最大值，第二個(gè)1代表只取最大值的索引
 
            #這兩行代碼是求準(zhǔn)確率的地方
            predicted = torch.max(output.data,1)[1]
            correct += (predicted == y_batch).sum()
            #print(correct)
            if batch_idx % 100 == 0:
                print('Epoch :{}[{}/{}({:.0f}%)]\t Loss:{:.6f}\t Accuracy:{:.3f}'.format(epoch,batch_idx * len(X_batch),len(train_loader.dataset),100.*batch_idx / len(train_loader),loss.data.item(),float(correct*100)/float(BATCH_SIZE)*(batch_idx+1)))
                
if __name__ == '__main__':
    myModel = LeNet5()
    print(myModel)
    train_model(myModel,train_loader)
    evaluate(myModel,test_loader,BATCH_SIZE)

epochs = 10
log_step_freq = 30
 
dfhistory = pd.DataFrame(columns = ['epoch','loss',metric_name,'val_loss','val_'+metric_name])
print('Start Training...')
nowtime = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')
print('========='*8 + '%s'%nowtime)
 
for epoch in range(1,epochs+1):
    
    #1.訓(xùn)練循環(huán)
    net.train()
    loss_sum = 0.0
    metric_sum = 0.0
    step = 1
    
    for step,(features,labels) in enumerate(dl_train,1):
        #梯度清零
        optimizer.zero_grad()
        
        #正向傳播求損失
        predictions = net(features)
        loss = loss_func(predictions,labels)
        metric = metric_func(predictions,labels)
        
        #反向傳播求梯度
        loss.backward()
        optimizer.step()
        
        #打印batch級(jí)別日志
        loss_sum += loss.item()
        metric_sum += metric.item()
        if step%log_step_freq == 0:
            print(('[Step = %d] loss: %.3f,' + metric_name+': %.3f %%')%(step,loss_sum/step,100*metric_sum/step))
            
    #2,驗(yàn)證循環(huán)
    net.eval()
    val_loss_sum = 0.0
    val_metric_sum = 0.0
    val_step =1
        
    for val_step,(features,labels) in enumerate(dl_valid,1):
        #關(guān)閉梯度計(jì)算
        with torch.no_grad():
            pred = net(features)
            val_loss = loss_func(pred,labels)
            val_metric = metric_func(labels,pred)
        val_loss_sum += val_loss.item()
        val_metric_sum += val_metric.item()
            
    #3,記錄日志
    info = (epoch,loss_sum/step,100*metric_sum/step,
            val_loss_sum/val_step,100*val_metric_sum/val_step)
    dfhistory.loc[epoch-1] = info
        
    #打印epoch級(jí)別日志
    print(('\nEPOCH = %d,loss = %.3f,' + metric_name+\
            '=%.3f %%,val_loss = %.3f'+' val_'+metric_name+'= %.3f %%')%info)
    nowtime = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')
    print('\n'+'=========='*8 + '%s'%nowtime)
print('Finishing Training...')    
 

def evaluteTop1(model, loader):
    model.eval()
    
    correct = 0
    total = len(loader.dataset)

    for x,y in loader:
        x,y = x.to(device), y.to(device)
        with torch.no_grad():
            logits = model(x)
            pred = logits.argmax(dim=1)
            correct += torch.eq(pred, y).sum().float().item()
        #correct += torch.eq(pred, y).sum().item()
    return correct / total

def evaluteTop5(model, loader):
    model.eval()
    correct = 0
    total = len(loader.dataset)
    for x, y in loader:
        x,y = x.to(device),y.to(device)
        with torch.no_grad():
            logits = model(x)
            maxk = max((1,5))
　　　　　　　 y_resize = y.view(-1,1)
            _, pred = logits.topk(maxk, 1, True, True)
            correct += torch.eq(pred, y_resize).sum().float().item()
    return correct / total

torch.topk(input, k, dim=None, largest=True, sorted=True, out=None) -> (Tensor, LongTensor)

import torch 
output = torch.tensor([[-5.4783, 0.2298],
                           [-4.2573, -0.4794],
                           [-0.1070, -5.1511],
                           [-0.1785, -4.3339]])

maxk = max((1,))  # 取top1準(zhǔn)確率，若取top1和top5準(zhǔn)確率改為max((1,5))
_, pred = output.topk(maxk, 1, True, True)

_
tensor([[ 0.2298],
        [-0.4794],
        [-0.1070],
        [-0.1785]])

pred
tensor([[1],
        [1],
        [0],
        [0]])