损失函数代码

这个是从代码层面，详细了解损失函数！

CrossEntropyLoss

 

import torch
import torch.nn as nn

# 实际中遇到的
outputs = torch.tensor([[0.5870, 0.4130],
                        [0.6517, 0.3483],
                        [0.4455, 0.5545],
                        [0.4786, 0.5214]], requires_grad=True)

y_train_tensor = torch.tensor([0, 0, 1, 1])

print(outputs)
print(y_train_tensor)
# 使用 CrossEntropyLoss 计算损失
criterion = nn.CrossEntropyLoss()
loss = criterion(outputs, y_train_tensor)
print('Outputs:', outputs)
print('Labels:', y_train_tensor)
print('Loss:', loss)

　　

 

,