audio

audio copied to clipboard

Added a Librispeech ctc recipe for demonstration purpose.

• This recipe demonstrates using either torch.nn.CTCLoss or k2.ctc_loss. Both can converge to similar results.
• It supports using either CTC or HMM topology with k2.ctc_loss.
• It currently uses a greedy decoder for decoding.

WER results will be posted later.

:link: Helpful Links

:test_tube: See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/audio/3541

• :page_facing_up: Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

:x: 11 New Failures

As of commit 54fdc286bd3a9a3155f8a51a9425fa1e79640545:

NEW FAILURES - The following jobs have failed:

• build (3.10, 11.8) / Build doc (gh)
• build (3.8) (gh)
• tests (3.10, 11.8) / linux-job (gh)
• tests (3.10) / linux-job (gh)
• tests (3.8, 11.8) / linux-job (gh)
• tests (3.8) / linux-job (gh)
• tests (3.9, 11.8) / linux-job (gh)
• tests (3.9) / linux-job (gh)
• tests / macos-job (gh)
• unittests-windows-cpu / windows-job (gh)
• unittests-windows-gpu / windows-job (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@huangruizhe has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

@huangruizhe has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

@huangruizhe has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.