使用Pytorch训练分类问题时,分类准确率的计算方式

#标签情况
print(y)
tensor([[1, 1, 0, 0]])
print(pred)
tensor([[1, 0, 1, 0]])
# 比较真实与预测
print(y==pred)
tensor([[ True, False, False,  True]])
# 对正确元素求和，sum会自动计算True的个数
print((y==pred).sum())
tensor(2)

for epoch in range(epochs):
    accuracy=0
    for i, (x,y) in enumerate(train_loader, 1): 
        pred = net(x)
        loss = loss_function(pred.to(torch.float32),y.to(torch.float32))
        optimizer.zero_grad()
        loss.backward() #反向传播
        optimizer.step() #更新梯度
        loss_steps[epoch]=loss.item()#保存loss
        running_loss = loss.item()
        accuracy += (pred == y).sum()
        acc = float(accuracy*100)/float(len(train_ids))# 除以元素总数，可以用其他方式获取
    print(f"第{epoch}次训练,loss={running_loss:.4f},Accuracy={acc:.3f}".format(epoch,running_loss,acc))

correct = torch.zeros(1).squeeze().cuda()
total = torch.zeros(1).squeeze().cuda()
for i, (images, labels) in enumerate(train_loader):
            images = Variable(images.cuda())
            labels = Variable(labels.cuda())
            output = model(images)
            prediction = torch.argmax(output, 1)
            correct += (prediction == labels).sum().float()
            total += len(labels)
acc_str = 'Accuracy: %f'%((correct/total).cpu().detach().data.numpy())

correct = list(0. for i in range(args.class_num))
total = list(0. for i in range(args.class_num))
for i, (images, labels) in enumerate(train_loader):
            images = Variable(images.cuda())
            labels = Variable(labels.cuda())
            output = model(images)
            prediction = torch.argmax(output, 1)
            res = prediction == labels
            for label_idx in range(len(labels)):
                label_single = label[label_idx]
                correct[label_single] += res[label_idx].item()
                total[label_single] += 1
 acc_str = 'Accuracy: %f'%(sum(correct)/sum(total))
 for acc_idx in range(len(train_class_correct)):
            try:
                acc = correct[acc_idx]/total[acc_idx]
            except:
                acc = 0
            finally:
                acc_str += '\tclassID:%d\tacc:%f\t'%(acc_idx+1, acc)