pytorch实现focal loss的两种方式小结

yipeiwu_com5年前Python基础

我就废话不多说了,直接上代码吧!

import torch
import torch.nn.functional as F
import numpy as np
from torch.autograd import Variable
'''
pytorch实现focal loss的两种方式(现在讨论的是基于分割任务)
在计算损失函数的过程中考虑到类别不平衡的问题,假设加上背景类别共有6个类别
'''
def compute_class_weights(histogram):
  classWeights = np.ones(6, dtype=np.float32)
  normHist = histogram / np.sum(histogram)
  for i in range(6):
    classWeights[i] = 1 / (np.log(1.10 + normHist[i]))
  return classWeights
def focal_loss_my(input,target):
  '''
  :param input: shape [batch_size,num_classes,H,W] 仅仅经过卷积操作后的输出,并没有经过任何激活函数的作用
  :param target: shape [batch_size,H,W]
  :return:
  '''
  n, c, h, w = input.size()

  target = target.long()
  input = input.transpose(1, 2).transpose(2, 3).contiguous().view(-1, c)
  target = target.contiguous().view(-1)

  number_0 = torch.sum(target == 0).item()
  number_1 = torch.sum(target == 1).item()
  number_2 = torch.sum(target == 2).item()
  number_3 = torch.sum(target == 3).item()
  number_4 = torch.sum(target == 4).item()
  number_5 = torch.sum(target == 5).item()

  frequency = torch.tensor((number_0, number_1, number_2, number_3, number_4, number_5), dtype=torch.float32)
  frequency = frequency.numpy()
  classWeights = compute_class_weights(frequency)
  '''
  根据当前给出的ground truth label计算出每个类别所占据的权重
  '''

  # weights=torch.from_numpy(classWeights).float().cuda()
  weights = torch.from_numpy(classWeights).float()
  focal_frequency = F.nll_loss(F.softmax(input, dim=1), target, reduction='none')
  '''
  上面一篇博文讲过
  F.nll_loss(torch.log(F.softmax(inputs, dim=1),target)的函数功能与F.cross_entropy相同
  可见F.nll_loss中实现了对于target的one-hot encoding编码功能,将其编码成与input shape相同的tensor
  然后与前面那一项(即F.nll_loss输入的第一项)进行 element-wise production
  相当于取出了 log(p_gt)即当前样本点被分类为正确类别的概率
  现在去掉取log的操作,相当于 focal_frequency shape [num_samples]
  即取出ground truth类别的概率数值,并取了负号
  '''

  focal_frequency += 1.0#shape [num_samples] 1-P(gt_classes)

  focal_frequency = torch.pow(focal_frequency, 2) # torch.Size([75])
  focal_frequency = focal_frequency.repeat(c, 1)
  '''
  进行repeat操作后,focal_frequency shape [num_classes,num_samples]
  '''
  focal_frequency = focal_frequency.transpose(1, 0)
  loss = F.nll_loss(focal_frequency * (torch.log(F.softmax(input, dim=1))), target, weight=None,
           reduction='elementwise_mean')
  return loss


def focal_loss_zhihu(input, target):
  '''
  :param input: 使用知乎上面大神给出的方案 https://zhuanlan.zhihu.com/p/28527749
  :param target:
  :return:
  '''
  n, c, h, w = input.size()

  target = target.long()
  inputs = input.transpose(1, 2).transpose(2, 3).contiguous().view(-1, c)
  target = target.contiguous().view(-1)

  N = inputs.size(0)
  C = inputs.size(1)

  number_0 = torch.sum(target == 0).item()
  number_1 = torch.sum(target == 1).item()
  number_2 = torch.sum(target == 2).item()
  number_3 = torch.sum(target == 3).item()
  number_4 = torch.sum(target == 4).item()
  number_5 = torch.sum(target == 5).item()

  frequency = torch.tensor((number_0, number_1, number_2, number_3, number_4, number_5), dtype=torch.float32)
  frequency = frequency.numpy()
  classWeights = compute_class_weights(frequency)

  weights = torch.from_numpy(classWeights).float()
  weights=weights[target.view(-1)]#这行代码非常重要

  gamma = 2

  P = F.softmax(inputs, dim=1)#shape [num_samples,num_classes]

  class_mask = inputs.data.new(N, C).fill_(0)
  class_mask = Variable(class_mask)
  ids = target.view(-1, 1)
  class_mask.scatter_(1, ids.data, 1.)#shape [num_samples,num_classes] one-hot encoding

  probs = (P * class_mask).sum(1).view(-1, 1)#shape [num_samples,]
  log_p = probs.log()

  print('in calculating batch_loss',weights.shape,probs.shape,log_p.shape)

  # batch_loss = -weights * (torch.pow((1 - probs), gamma)) * log_p
  batch_loss = -(torch.pow((1 - probs), gamma)) * log_p

  print(batch_loss.shape)

  loss = batch_loss.mean()
  return loss

if __name__=='__main__':
  pred=torch.rand((2,6,5,5))
  y=torch.from_numpy(np.random.randint(0,6,(2,5,5)))
  loss1=focal_loss_my(pred,y)
  loss2=focal_loss_zhihu(pred,y)

  print('loss1',loss1)
  print('loss2', loss2)
'''
in calculating batch_loss torch.Size([50]) torch.Size([50, 1]) torch.Size([50, 1])
torch.Size([50, 1])
loss1 tensor(1.3166)
loss2 tensor(1.3166)
'''

以上这篇pytorch实现focal loss的两种方式小结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【听图阁-专注于Python设计】。

相关文章

python自动发微信监控报警

python自动发微信监控报警

我们每个人每天都是在用微信,在程序开发过程中,我们会需要监控我们的程序,发短信监控收费,发邮件懒得看,发微信是最好的方式,而且是免费的。发现个非常好用的python库:wxpy。wxpy...

python实现生成Word、docx文件的方法分析

本文实例讲述了python实现生成Word、docx文件的方法。分享给大家供大家参考,具体如下: http://python-docx.readthedocs.io/en/latest/...

Python中max函数用于二维列表的实例

最近写一个和二维列表有关的算法时候发现的 当用max求二维列表中最大值时,输出的结果是子列表首元素最大的那个列表 测试如下 c=[[1,2,-1],[0,5,6]] a=[[0,3...

selenium+python自动化测试之页面元素定位

selenium+python自动化测试之页面元素定位

上一篇博客selenium+python自动化测试(二)–使用webdriver操作浏览器讲解了使用webdriver操作浏览器的各种方法,可以实现对浏览器进行操作了,接下来就是对浏览器...

如何解决django配置settings时遇到Could not import settings 'conf.local'

举个例子吧 Django最佳实践与部署:Nginx + Gunicorn + Supervisor(Ubuntu和CentOS) http://sfdye.com/articles/dj...