GitLab

fix whl package naming

See merge request !4

specify CMake version

See merge request !3

fix 104994 and add jenkins.sh

See merge request !2

(cherry picked from commit 994de774)

Summary:
Pull Request resolved: https://github.com/pytorch/audio/pull/3803

The model checkpoint path can not be created for Squim models. Use the latest download_asset method to fix it.

Reviewed By: moto-meta

Differential Revision: D59061348

Differential Revision: D57294298

Pull Request resolved: https://github.com/pytorch/audio/pull/3791

[codemod] c10::optional -> std::optional in pyspeech/experimental/csrc/decoders/TransducerDecoder.h +20

Differential Revision: D57294284

Pull Request resolved: https://github.com/pytorch/audio/pull/3793

[codemod] c10::optional -> std::optional in pytorch/audio/src/libtorio/ffmpeg/stream_reader/stream_processor.h +20

Differential Revision: D57294285

Pull Request resolved: https://github.com/pytorch/audio/pull/3792

Unpin mkl version and install pybind11 to get the windows CI working again

This fixes #3767

* Update tacotron2_pipeline_tutorial.py

- Fixed typo
- Clarified what was being done in different sections

Differential Revision: D54263224

Pull Request resolved: https://github.com/pytorch/audio/pull/3751

The lengths of targets and log_probs should be reversed.

Differential Revision: D53606067

Pull Request resolved: https://github.com/pytorch/audio/pull/3740

* add golf and dynonet paper

* doc: add references

* add EOF

* fix: line too long

* remove line end space

* remove indentation
Co-authored-by: moto <855818+mthrok@users.noreply.github.com>

---------
Co-authored-by: moto <855818+mthrok@users.noreply.github.com>

Summary:
This is not needed anymore after https://github.com/pytorch/test-infra/pull/4865.


Reviewed By: malfet, jeanschmidt, clee2000, NicolasHug

Differential Revision: D52735187

Pulled By: huydhn

* Update doc

* Update citation

Addresses #3688 

Back port from release/2.1 branch.
Need to git-fetch source code to get the version number dynamically

global audio backend is removed thus this is no-op.

PyTorch lightening is having issue with the nightly PyTorch.
Let the other tests still run.