Default hyper parameter settings of lr_find

If anyone could guide me through the lr_find’s default parameters and does it use one cycle policy in itself?

send me the code that you used.

lr_find and its default parameters are documented pretty well here. The only default param that is non-obvious is the wd (weight decay) parameter, which you can see the default for here (it is 1e-2). Are there any more specific questions you have about them?

Also, lr_find doesn’t use the one cycle policy - you use the results from it to set your maximum learning rate for the one cycle policy. That part is documented here.

Got it.

I’m looking for some details about this.