pytorch的交叉熵损失函数计算细节CrossEntropyLoss()

返回值=(各张图片的损失值之和/图片数量)，也就是各个图片的损失求平均
 

以下代码模拟计算第一张图片的损失

x = torch.tensor([0.1, 0.2, 0.3])
y = torch.tensor([1])
x = torch.reshape(x, (1, 3))
loss_cross = nn.CrossEntropyLoss()
result_cross = loss_cross(x, y)
print(result_cross)

以下代码模拟计算第二张图片的损失
x = torch.tensor([3.2, 0.6, 0.9])
y = torch.tensor([1])
x = torch.reshape(x, (1, 3))
loss_cross = nn.CrossEntropyLoss()
result_cross = loss_cross(x, y)
print(result_cross)

以下代码模拟计算两张图片一起求loss的损失

x = torch.tensor([[0.1, 0.2, 0.3],
                  [3.2, 0.6, 0.9]])
y = torch.tensor([1,1])
#x = torch.reshape(x, (2, 3))
loss_cross = nn.CrossEntropyLoss()
result_cross = loss_cross(x, y)
print(result_cross)

三个损失值分别为

tensor(1.1019)
tensor(2.7609)
tensor(1.9314)