Steven Atkinson

@kf-rahman go for it! I'd prefer if you do it by extending the checkpoint class and wrap the save action with some try/except and retry a few times. Sound good?...

This sounds very helpful :) I'm going to move this over to the training repo where this is calculated. > it would require some experimental tweaking to maintain the -18dB(RMS)...

Related: https://github.com/sdatkinson/NeuralAmpModelerCore/issues/159#issuecomment-3239702643

TLDR: I think you're confusing NAM's "basic" trainer with NAM itself. > it is not recommended (or even possible?) to create nam models that contain time based effects (chorus, delay,...

Now that we'll be only supporting v3_0_0.wav (#443) this can be streamlined; updating Issue.

Hmm. I've seen this on Colab. I wasn't able to reproduce it just now on my desktop with my "day to day" nam environment, but I just made a new...

Graceful shutdown appears to attempt to `exit(1)` at the end of its `except` block in version 2.5.0.post0. From [trainer/call.py](https://github.com/Lightning-AI/pytorch-lightning/blob/cd785a4ba437e0549646fede6c1bf515d1a39124/src/lightning/pytorch/trainer/call.py#L35-L72): ```python def _call_and_handle_interrupt(trainer: "pl.Trainer", trainer_fn: Callable, *args: Any, **kwargs: Any) ->...

@DomMcsweeney @FranOjedaCMusic not to worry, I can transfer the Issue across repos (thanks, GitHub!) @FranOjedaCMusic , can you have a look at the Issue template for bug reports on this...

Can you please provide the rest of the information that the Bug Report template requests?

I mean the rest of the questions the bug template asks for. Does this only happen when a NAM is loaded? If so, which one? etc. Please answer this as...