Pytorch 实现计算分类器准确率(总分类及子分类)

时间:2021-07-22 04:12:17

分类器平均准确率计算:

?
1
2
3
4
5
6
7
8
9
10
11
12
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())

分类器各个子类准确率计算:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
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)

以上这篇Pytorch 实现计算分类器准确率(总分类及子分类)就是小编分享给大家的全部内容了,希望能给大家一个参考,也希望大家多多支持服务器之家。

原文链接:https://blog.csdn.net/u014657795/article/details/86419197