GitLab

Porting Tacotron2 from https://github.com/NVIDIA/DeepLearningExamples/blob/master/PyTorch/SpeechSynthesis/Tacotron2/tacotron2/model.py

`torchaudio._internal.fft` was originally added to account for the introduction of module `torch.fft`, when `torch.fft` could refer to either a module or function. Now that `torch.fft` refers unambiguously to a module, we remove `torchaudio._internal.fft` and replace references to it with `torch.fft`.

The test was broken from the beginning and for a reason unrelated to
ROCm, but was disabled for ROCm in https://github.com/pytorch/audio/pull/1411.
https://github.com/pytorch/audio/pull/1604 fixed the test so that we are re-enabling it.

* [BC-Breaking] Default to PCM_16 for flac on soundfile backend

Resolving https://github.com/pytorch/audio/issues/1592

The test backend/soundfile/save_test.py::TestFileObject::test_fileobj_flac
is failing due to the fact that when soundfile.write received
subtype=None (for flac files), it would fall back to 'PCM_16'.
But in the same time, torchaudio will use 'PCM_24', which would
result in distorted signal.

This commit fixed this issue by changing the default bit-per-sample
of for flac files from 24-bit to 16-bit.

No longer needed for Python 3.9
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>

Needed to allow building on M1

Downloaded from https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD and https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD

Following https://github.com/pytorch/pytorch/blob/master/docs/cpp/source/notes/inference_mode.rst#migration-guide-from-autononvariabletypemode 

Resolves #1522 

There are some issues on CMake-based integration after 2021-06-19.

```
  Imported target "torch" includes non-existent path
    "/opt/conda/conda-bld/pytorch_1624259172741/work/torch/lib"
  in its INTERFACE_INCLUDE_DIRECTORIES.  Possible reasons include:
  * The path was deleted, renamed, or moved to another location.
  * An install or uninstall procedure did not complete successfully.
  * The installation package was faulty and references files it does not provide.
```

This is being worked on PR like https://github.com/pytorch/pytorch/pull/60403.
The nightly builds are still causing the error, so as a workaround, 
we pin the nightly build version for Linux/macOS unittest.
The issue still happens on Windows unittest and binary builds.

Mel scale frequency is in general defined/used on power spectrogram and as far as we know, it is not defined for raw (complex value) spectrogram. When `power=None` (when intermediate spectrogram is complex-valued), `MelSpectrogram` fails because of extra dimension (when using pseudo complex) or type mismatch between `torch.float` and torch.cfloat` (when using native complex).

This commit changes type annotation so that power is only float.

Since 0.9.0-RC1, `T.Resample` precomputes and caches resampling kernel for performance improvement. (10x improvement).

The implementation from 0.8.0 computed the kernel on-the-fly on the same `device`/`dtype` as the input Tensor, 
but in the newer version, the kernel is precomputed at the construction time and is cached with `float32` first.
This causes degradation if one wants to perform resampling on `float64`, because `sinc` values computed on `float32`s are not good enough for resampling in `float64`.

The reason why we decided to use `float32` for initial caching is to keep the UX disruption minimum, and there were no way to make it work for `float64`. This PR adds `dtype` argument, that can be used for overwriting the cache precision.

`torchaudio.compliance.kaldi.resample_waveform` has been replaced with `torchaudio.funcitonal.resample`.