Sevag H

By my method, do you mean PMPM/Probablistic McLeod Pitch Method?

It's a good idea. I think some scripts around the fetching, loading, running pitch tracking, and evaluating the quality of the results on the NSynth data would be a good...

As long as it's a configurable or well-known path e.g. `~/NSYNTH_DATA`, that makes sense to me.

You can do it pretty easily with the included Docker image. 1. Use the standard `checkout` workflow of github actions (https://github.com/actions/checkout) to clone the code (make sure you use the...

Feel free to explore at your pace, and ping me when you have any questions. There is no urgency on this issue.

Did you make sure to enable git-lfs? >Use the standard checkout workflow of github actions (https://github.com/actions/checkout) to clone the code (make sure you use the appropriate options to enable git-lfs,...

Also I tested with the numpy code (i.e. same tensors with `.detach().cpu().numpy()` and got the same outcome: ``` File "/usr/local/lib/python3.8/dist-packages/numpy/linalg/linalg.py", line 88, in _raise_linalgerror_singular raise LinAlgError("Singular matrix") numpy.linalg.LinAlgError: Singular matrix...

I just had a clue; since I know I use auraloss.SISDRLoss() (Scale-Invariant SDR loss, https://github.com/csteinmetz1/auraloss) with success in my neural network, I tried your `si_sdr_loss` function, where I don't get...

Thanks for the suggestion, if I use a large enough value e.g. `load_diag=1e-5`, I get past this issue. Indeed, basically in the early iterations of my neural network, there is...

@cbeuw btw, I was able to use this PR as a BLAS library in a C++ project (only sgemm/sgemv, a small subset, but a real usecase nonetheless). It uses Eigen...