Closed JackFurby closed 4 years ago
Hi @JackFurby, thanks for raising this issue. Previously, my package did not consider the data parallelism case.
I think I have figured out how to add data parallelism support, but I do not have any experience in doing it, I think I should upload a beta package for you to test out first.
You should remove existing nnAudio installation by pip uninstall nnAudio
Then install the beta package by pip install nnAudio==0.0.11b
I did some modifications to Spectrogram.MelSpectrogram only, can you test it out if it works for you before I do a complete update on my package?
I am trying to use this library with multiple GPU's but am getting the following error message:
I have looked into the point in which the code is stopping which seems to be when the batch is split across multiple GPUs and passed through the model. I believe this is because when I am initialising the model, I am configuring the MelSpectrogram with
device=device
. For a single GPU or a CPU this is fine, but moving over to multiple GPUs, this is being fixed to just one of the GPUs. I am not sure if the issue lies with my configuration or with the library itself, but I am after a way of having the device set on the fly.My model implementation is as follows: