GitLab

- TorchScript-able `Wav2Vec2Model` class
- Factory functions for three configurations presented in the paper 
  - `wav2vec2_base`
  - `wav2vec2_large`
  - `wav2vec2_large_lv60k`

* Remove `class FunctionalComplex` header accidentally re-introduced in #1490 

* Update VAD docstring and check for input shape length

* Update docstring in forward for transform

* Address review feedback: merge tests, update wording

Since the update of Xcode, some of HTTP tests are failing on macOS, due to `reqests.get` fails to establish a connection to local server. This change increase the wait time to 2 seconds before test starts.

Co-authored-by: Vincent Quenneville-Belair <vincentqb@gmail.com>

It was reported in #1478 that spectrogram masking operations were done in-place and modified the original input tensors. This PR fixes this behavior and adds tests to ensure that the input tensor is not changed.

Replace the prototype RNNT implementation (using warp-transducer) with one without external library dependencies

- Put functional test logic into one place, `functional_impl.py`
- Tidy imports

* Fixed floor_divide deprecation warnings seen in pytest output

* Fixed warning in test_flanger_triangle_linear

The VAD function trims the input tensor to the first instance of voice
activity on any channel or item. Trimming batches this way may be
undesirable as the item with earliest activity will dominate. Either
way, the batch behaviour does not match the itemwise behaviour.

The VAD batch consistency tests currently pass out of coincidence, but
they specify incorrect behaviour. This commit removes them.

Use shorter input sequences to avoid time out error on CI

* Remove an invalid InverseMel comparison unit test

Similar to #1426 `test_InverseMelScale` in `librosa_compatibility_test` is not
ensuring the comaptibility to librosa. Having this test can give a wrong statement
about the librosa numerical compatibility about the function.

* Add test for InverseMelScale

The new test compares the result of inverse mel scale against the reference spectrogram, so no need to use librosa.
This test serves as more like an insurance that the change to the implementation of InverseMelScale only improves the result, not the other way.