Update torch support

[Fhrozen]

This pull request updates the GitHub Actions CI workflows to support newer versions of torchaudio and expands the workflow triggers to run on all branches, not just main. The changes also update the Python version matrix and simplify the configuration.

Workflow trigger updates:

• Both .github/workflows/ci.yml and .github/workflows/espnet.yml now trigger on pushes to any branch, not just main, in addition to pull requests. [1] [2]

Dependency and matrix updates:

• The torchaudio-version matrix in both workflows is updated to test only recent versions: 2.5.1, 2.7.1, 2.8.0, and 2.9.1, removing older versions. [1] [2]
• The Python version matrix now explicitly includes Python 3.12 with the latest torchaudio version (2.9.1). [1] [2]
• The include section is simplified to only add the Python 3.12 and torchaudio 2.9.1 combination. [1] [2]

[gemini-code-assist]

Note

Gemini is unable to generate a review for this pull request due to the file types involved not being currently supported.

[Fhrozen]

/gemini review

[gemini-code-assist]

Code Review

This pull request updates the project's dependencies, specifically bumping torch and torchaudio to version 2.5.1 and removing legacy audio backend configurations and unused imports. However, the tox.ini file includes several non-existent torchaudio versions (2.7.1, 2.8.0, and 2.9.1) which will cause environment setup failures.

[gemini-code-assist]

The default environment audio2.9.1 refers to a version of torchaudio that does not exist on PyPI. This will cause tox to fail by default. Based on current PyTorch releases, the latest stable version is 2.5.1.

envlist = common_upstream-py311-audio2.5.1,

[gemini-code-assist]

The torchaudio versions 2.7.1, 2.8.0, and 2.9.1 are not available on PyPI. torchaudio versions typically track torch versions, which are currently at 2.5.x. Specifying these non-existent versions will cause CI failures. Please use valid versions such as 2.2.2, 2.3.1, 2.4.1, or 2.5.1.