GitLab

* add basics for Fade

* add fade possibilities: at start, end or both

* add different types of fade

* add docstrings, add overriding possibility

* remove unnecessary logic

* correct typing

* agnostic to batch size or n_channels

* add batch test to Fade

* add transform to options

* add test_script_module

* add coherency with test batch

* remove extra step for waveform_length

* update docstring

* add test to compare fade with sox

* change name of fade_shape

* update test fade vs sox with new nomenclature for fade_shape

* add Documentation
Co-authored-by: Vincent QB <vincentqb@users.noreply.github.com>

* change default value of dither.

* update doc.

* get typing on Docstrings right

* Improve Documentation standardise

* phase_advance should be a buffer so it moves device correctly

* flake8
Co-authored-by: Vincent QB <vincentqb@users.noreply.github.com>

* AmplitudeToDB to torch.nn.Module

* TimeStretch use torch.nn.Module

* Inverse Mel Scale Implementation

* Inverse Mel Scale Docs

* Better working version.

* GPU fix

* These shouldn't go on git..

* Even better one, but does not support JITability.

* Remove JITability test

* Flake8

* n_stft is a must

* minor clean up of initialization

* Add librosa consistency test

This PR follows up #366 and adds test for `InverseMelScale` (and `MelScale`) for librosa compatibility.

For `MelScale` compatibility test;
1. Generate spectrogram
2. Feed the spectrogram to `torchaudio.transforms.MelScale` instance
3. Feed the spectrogram to `librosa.feature.melspectrogram` function.
4. Compare the result from 2 and 3 elementwise.
Element-wise numerical comparison is possible because under the hood their implementations use the same algorith.

For `InverseMelScale` compatibility test, it is more elaborated than that.
1. Generate the original spectrogram
2. Convert the original spectrogram to Mel scale using `torchaudio.transforms.MelScale` instance
3. Reconstruct spectrogram using torchaudio implementation
3.1. Feed the Mel spectrogram to `torchaudio.transforms.InverseMelScale` instance and get reconstructed spectrogram.
3.2. Compute the sum of element-wise P1 distance of the original spectrogram and that from 3.1.
4. Reconstruct spectrogram using librosa
4.1. Feed the Mel spectrogram to `librosa.feature.inverse.mel_to_stft` function and get reconstructed spectrogram.
4.2. Compute the sum of element-wise P1 distance of the original spectrogram and that from 4.1. (this is the reference.)
5. Check that resulting P1 distance are in a roughly same value range.

Element-wise numerical comparison is not possible due to the difference algorithms used to compute the inverse. The reconstructed spectrograms can have some values vary in magnitude.
Therefore the strategy here is to check that P1 distance (reconstruction loss) is not that different from the value obtained using `librosa`. For this purpose, threshold was empirically chosen

```
print('p1 dist (orig <-> ta):', torch.dist(spec_orig, spec_ta, p=1))
print('p1 dist (orig <-> lr):', torch.dist(spec_orig, spec_lr, p=1))
>>> p1 dist (orig <-> ta): tensor(1482.1917)
>>> p1 dist (orig <-> lr): tensor(1420.7103)
```

This value can vary based on the length and the kind of the signal being processed, so it was handpicked.

* Address review feedbacks

* Support arbitrary batch dimensions.

* Add batch test

* Use view for batch

* fix sgd

* Use negative indices and update docstring

* Update threshold
Co-authored-by: Charles J.Y. Yoon <jaeyeun97@gmail.com>

* add speechcommand dataset and test

* prepend the full path to each result

* add missing param on docstring in walk_files

* add file to run tests on SpeechCommand Dataset

* reduce logic

* update test on SpeechCommands

* correct the indentation on docstring walk_files

* flake8 compliance

* change tuple type returned. move path split logic in load item.

* typo in name.

* redundant file path.

* filter background noise.
Co-authored-by: Vincent QB <vincentqb@users.noreply.github.com>

* LJ Speech dataset

* refactoring

as per @vincentqb's suggestions

* workaround for bartlett_window https://github.com/pytorch/pytorch/issues/32358#issuecomment-576909755

* only change dtype.

Signed-off-by: Eli Uriegas <eliuriegas@fb.com>

This reverts commit cdf5c83d.

Python 2.7 was EOL on January 1, 2020

The last torchaudio release to support Python 2.7 was 0.4.0
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>

CUDA_SUFFIX was still being used here when it should've been swapped out
for PYTORCH_VERSION_SUFFIX, which is what's being used for conda below.
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>
(cherry picked from commit 009b115d074ac5fcca2cc34662fe814df63324c1)
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>

* extend batch support

closes #383

* function for batch test.

* set seed.

* adjust tolerance for griffinlim.

* fix power of spectrogram. makes power a float.

closes #389

* commenting out failing test.

* change skip test logic for librosa.
closes #373

* move test for scriptmodule.

* avoiding code duplication.

With the introduction of the `filter_branch` parameter to the `workflows`
function we no longer have a need to have this if block anymore per
@ezyang's assessment.
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>

* [WIP] Add Windows CI

* Remove cu_version

* checkout_merge -> checkout

* Add build script

* Switch backend to soundfile

* Remove soundfile as dependency

* Rename jobs

* Fix lint

closes #387

closes #382

Remove suspect logic.

Closes #304

See rationale writeup: https://github.com/pytorch/vision/pull/1321#issuecomment-531033978

* Declare file encoding to support special characters

* fix missing utf_8_encoder error in Travis tests

* Py 2.7 backwards-compat iterator

* ensure integer argument to torch.nn.functional.pad

* cast match.ceil result as integer

* test with jit.

* test passed after adding annotation, and removing get_default_dtype

* fix conversion error.

* moving test to transform.

* reverting to original test.

* move type.

* math.gcd added in python 3.5.
Co-authored-by: Vincent QB <vincentqb@users.noreply.github.com>

* Griffin-Lim Transformation Implementation

* Griffin-Lim Docs

* Remove f-string from backwards compatibility

* iSTFT is now jit-able.

* Comment changes

* Functional Implementation & now jitable

* flake8

* Doc & GPU Fix

* Librosa comparison test

* test directly griffinlim's output. tighter atol.

* matching signature to docstring.
Co-authored-by: Vincent QB <vincentqb@users.noreply.github.com>

* Fixed GPU tests

* Simplify lfilter functional

* use `torch.clamp` instead of `torch.min(..., torch.max(...))`
* remove unneeded creation of ones tensor for previous method

The current lfilter function uses min and max to essentially do a clamp
function.  I changed the code to use clamp instead.  It is more readable
than the previous version.

FYI, if you want to keep the previous way, you could make a
broadcastable tensor of size 1 instead of creating a tensor the size of
the input.
Signed-off-by: David Pollack <david@da3.net>

* Parallelize waveform windows calculation

I've parallelized the calculation of the waveform windows and also
removed the inefficient calculation within the for-loop.
Signed-off-by: David Pollack <david@da3.net>

* Refactoring and minor readability changes
Signed-off-by: David Pollack <david@da3.net>

* Remove one more creation of a temporary tensor
Signed-off-by: David Pollack <david@da3.net>

* move sox inside function calls.

* add backend switch mechanism.

* import sox at runtime, not import.

* add backend list.

* backend tests.

* creating hidden modules for backend.

* naming backend same as file: soundfile.

* remove docstring in backend file.

* test soundfile info.

* soundfile doesn't support int64.

* adding test for wav file.

* error with incorrect parameter instead of silent ignore.

* adding test across backend. using float32 as done in sox.

* backend guard decorator.

* Fix MelScale test and documentation

* revert change to tests