Iver Jordal

Thanks RoyJames :) Just so I understand your way of using torch-audiomentations, I'd like to know: Did you run the transforms on CPU (in each data loader worker)? And did...

I should write [this article](https://github.com/iver56/audiomentations/issues/230) soon, to make it easier to decide if torch-audiomentations is a good fit and how. Also, it would be swell if someone/we could reproduce and...

> This way I essentially define the GPU preprocessor function as part of my trainer (rather than the dataloader), and call it first in the forward() function after each mini-batch...

Interesting!

You were the one who suggested it 😉

Sure, we can get by without being rigid about the docstring format. Docstring format is not crucial right now

Good! Yeah, I agree, let's get rid of the librosa dependency. I haven't followed the progress in torchaudio closely. Did they speed up their resampling since https://github.com/asteroid-team/torch-audiomentations/issues/37#issuecomment-724691882 ?

Thanks for the links :) Those would probably be a good fit for [audiomentations](https://github.com/iver56/audiomentations/issues/11)

I'm not so fond of torchaudio's resample function, because it seems to be much slower than julius. Here's the result of a crude benchmark that resamples some audio from 44100...

In my crude benchmark, I ran it simply like this: ``` import torch from torchaudio.compliance.kaldi import resample_waveform for lowpass_filter_width in (2, 6, 10): with timer("pytorch-audio kaldi-compliant LPF width={}".format(lowpass_filter_width)): pytorch_kaldi_compliant =...