pytorch怎么使用交叉熵損失函數(shù)？weight參數(shù)的使用介紹！

猿友 2021-07-16 10:18:56 瀏覽數(shù) (4592)

反饋

在pytorch的交叉熵損失函數(shù)的學習中，weight慘啊作為交叉熵函數(shù)對應參數(shù)的輸入值，它的使用并不是想象中的那么簡單。接下來的這篇文章小編就來詳細的介紹一下交叉熵損失函數(shù)的weight參數(shù)怎么使用吧。

首先

必須將權重也轉為Tensor的cuda格式；

然后

將該class_weight作為交叉熵函數(shù)對應參數(shù)的輸入值。

補充：關于pytorch的CrossEntropyLoss的weight參數(shù)

首先這個weight參數(shù)比想象中的要考慮的多

你可以試試下面代碼

import torch
import torch.nn as nn
inputs = torch.FloatTensor([0,1,0,0,0,1])
outputs = torch.LongTensor([0,1])
inputs = inputs.view((1,3,2))
outputs = outputs.view((1,2))
weight_CE = torch.FloatTensor([1,1,1])
ce = nn.CrossEntropyLoss(ignore_index=255,weight=weight_CE)
loss = ce(inputs,outputs)
print(loss)

tensor(1.4803)

這里的手動計算是：

loss1 = 0 + ln(e0 + e0 + e0) = 1.098

loss2 = 0 + ln(e1 + e0 + e1) = 1.86

求平均 = (loss1 *1 + loss2 *1)/ 2 = 1.4803

加權呢？

import torch
import torch.nn as nn
inputs = torch.FloatTensor([0,1,0,0,0,1])
outputs = torch.LongTensor([0,1])
inputs = inputs.view((1,3,2))
outputs = outputs.view((1,2))
weight_CE = torch.FloatTensor([1,2,3])
ce = nn.CrossEntropyLoss(ignore_index=255,weight=weight_CE)
loss = ce(inputs,outputs)
print(loss)

tensor(1.6075)

手算發(fā)現(xiàn)，并不是單純的那權重相乘：

loss1 = 0 + ln(e0 + e0 + e0) = 1.098

loss2 = 0 + ln(e1 + e0 + e1) = 1.86

求平均 = (loss1 * 1 + loss2 * 2)/ 2 = 2.4113

而是

loss1 = 0 + ln(e0 + e0 + e0) = 1.098

loss2 = 0 + ln(e1 + e0 + e1) = 1.86

求平均 = (loss1 *1 + loss2 *2) / 3 = 1.6075

發(fā)現(xiàn)了么，加權后，除以的是權重的和，不是數(shù)目的和。

我們再驗證一遍：

import torch
import torch.nn as nn
inputs = torch.FloatTensor([0,1,2,0,0,0,0,0,0,1,0,0.5])
outputs = torch.LongTensor([0,1,2,2])
inputs = inputs.view((1,3,4))
outputs = outputs.view((1,4))
weight_CE = torch.FloatTensor([1,2,3])
ce = nn.CrossEntropyLoss(weight=weight_CE)
# ce = nn.CrossEntropyLoss(ignore_index=255)
loss = ce(inputs,outputs)
print(loss)

tensor(1.5472)

手算：

loss1 = 0 + ln(e0 + e0 + e0) = 1.098

loss2 = 0 + ln(e1 + e0 + e1) = 1.86

loss3 = 0 + ln(e2 + e0 + e0) = 2.2395

loss4 = -0.5 + ln(e0.5 + e0 + e0) = 0.7943

求平均 = (loss1 * 1 + loss2 * 2+loss3 * 3+loss4 * 3) / 9 = 1.5472

可能有人對loss的CE計算過程有疑問，我這里細致寫寫交叉熵的計算過程，就拿最后一個例子的loss4的計算說明