loss_weights decay with the trainning process?

[Xhao12138]

I want to achieve the loss weight as the training process goes down.This is my code:
model.complie('adam', lr=0.001, loss_weights=[1,10000], external_trainable_variables=C)
I hope 10000 to linearly decay to 1 when the trainning is over.
Thanks for your help!

[lululxvi]

There is no decay in DeepXDE.