Commit Graph

19 Commits

Author SHA1 Message Date
Matthew Boakes 1b9c400bca
PyTorch 2.1 Updates (Weight Norm and TorchAudio I/O) (#3176)
* Replaced PyTorch weight_norm With parametrizations.weight_norm

* TorchAudio: Migrating The I/O Functions To Use The Dispatcher Mechanism

* Corrected Code Style

---------

Co-authored-by: Eren Gölge <erogol@hotmail.com>
2023-11-09 16:31:03 +01:00
Martin Weinelt 994be163e1
Use packaging.version for version comparisons (#2310)
* Use packaging.version for version comparisons

The distutils package is deprecated¹ and relies on PEP 386² version
comparisons, which have been superseded by PEP 440³ which is implemented
through the packaging module.

With more recent distutils versions, provided through setuptools
vendoring, we are seeing the following exception during version
comparisons:

> TypeError: '<' not supported between instances of 'str' and 'int'

This is fixed by this migration.

[1] https://docs.python.org/3/library/distutils.html
[2] https://peps.python.org/pep-0386/
[3] https://peps.python.org/pep-0440/

* Improve espeak version detection robustness

On many modern systems espeak is just a symlink to espeak-ng. In that
case looking for the 3rd word in the version output will break the
version comparison, when it finds `text-to-speech:`, instead of a proper
version.

This will not break during runtime, where espeak-ng would be
prioritized, but the phonemizer and tokenizer tests force the backend
to `espeak`, which exhibits this breakage.

This improves the version detection by simply looking for the version
after the "text-to-speech:" token.

* Replace distuils.copy_tree with shutil.copytree

The distutils module is deprecated and slated for removal in Python
3.12. Its usage should be replaced, in this case by a compatible method
from shutil.
2023-01-29 23:47:00 +01:00
p0p4k 903a77c197
Update wavenet.py (#1796)
* Update wavenet.py

Current version does not use "in_channels" argument. 
In glowTTS, we use normalizing flows and so "input dim" == "ouput dim" (channels and length). So, the existing code just uses hidden_channel sized tensor as input to first layer as well as outputs hidden_channel sized tensor. 
However, since it is a generic implementation, I believe it is better to update it for a more general use.

* "in_channels -> hidden_channels"
2022-08-01 12:20:37 +02:00
Eren Gölge fd95926009 Update GlowTTS 2021-09-30 14:47:56 +00:00
Eren Gölge c312acac7d Implement VITS model 🚀
VITS model implementation built on Glow TTS and HiFiGAN
layers.
2021-08-09 18:02:36 +00:00
Eren Gölge 2e1a428b83 Update glowtts docstrings and docs 2021-06-30 14:30:55 +02:00
Eren Gölge d42d1c02ea Use `torch.linalg.qr` for pytorch > `v1.9.0` 2021-06-28 17:03:47 +02:00
Eren Gölge f519012dea reformatting and styling 2021-04-12 11:47:39 +02:00
root 5c87753e88 glow-tts fix for saving inverse weight 2021-01-20 02:09:42 +00:00
erogol b206162d11 more docstrings 2021-01-11 17:25:04 +01:00
erogol de2a542f83 glow-tts bug fix 2021-01-07 13:40:32 +01:00
erogol d3b7284be4 glow-tts comments and refactoring 2021-01-06 13:19:40 +01:00
erogol e82d31b6ac glow ttss refactoring 2021-01-06 13:19:40 +01:00
erogol 4ef083f0f1 select decoder type for SS 2021-01-06 13:19:40 +01:00
erogol 7b20d8cbd3 implement residual BN convolution and add it as an alternative encoder for glow-tts. also generic layers to layers/generic 2021-01-06 13:19:40 +01:00
erogol 665f7ca714 linter fix 2020-09-24 12:57:54 +02:00
erogol 3660c57f1e time seperable convolution encoder, huber loss for duration predictor 2020-09-17 03:10:58 +02:00
erogol 45fbc0d003 convolution encoder with GLU and res connections 2020-09-12 03:40:21 +02:00
erogol 15e6ab3912 glow-tts module renaming updates 2020-09-12 03:33:36 +02:00