xùn liàn / shén jīng wǎng luò / shí / wǒ men / cháng yòng / jiāo chā shāng / zuò wéi / sǔn shī / hán shù

训练神经网络时，我们常用交叉熵作为损失函数。

When training a neural network, we often use cross-entropy as the loss function.