Difference between MultiplicativeLR and LambdaLR

Question

Can someone explain difference between torch.optim.lr_scheduler.LambdaLR(https://pytorch.org/docs/stable/optim.html) and torch.optim.lr_scheduler.MultiplicativeLR(https://pytorch.org/docs/stable/optim.html)?

Here is brief description of MultiplicativeLR:

and LambdaLR

Answer 1

The main difference is that they use a different function for computing the learning rate of the function.

LambdaLR's function is:

While MultiplicativeLR's function is:

Thus they would have a different result for the learning rate. Both are useful for a variety of scenarios.