confusion in the loss function

[Flora279]

Hi,
I am going through your code, however, I cannot understand this line of code:
loss1 = args.weight_diff * K.mean(model1.get_layer('prediction').output)
why we are taking derivative wrt this loss? Shouldn't it be the M0 (diff in the predictions) in the paper?

Thanks.

[lizhenghui918]

嗨， 我正在查看您的代码，但是，我无法理解这行代码： loss1 = args.weight_diff * K.mean(model1.get_layer('prediction').output) 为什么我们要对这种损失进行导数？不应该是论文中的 M0（预测中的差异）吗？

谢谢。

你好，我目前也是在看这段代码，请问你的代码能够运行吗？我的代码一直在driving-modle文件报错，请问您能够帮助我吗，非常感谢
Hello, I'm also looking at this code at present. Can your code run? My code has been reporting errors in the driving mode file. Can you help me