GitLab

Summary:
following pr https://github.com/pytorch/audio/issues/2716
- For preprocessing
  - The HuBERT feature takes lots of memory which may not fit some machines. Enable to use a subset of feature for training a k-means model.

- For pre-training
  - Normalize the loss based on the total number of masked frames across all GPUs.
  - Use mixed precision training. fp16 is not well supported in pytorch_lightning.
  - Log accuracies of masked/unmasked frames during training.
  - Clip the gradients with norm `10.0`.

- For ASR fine-tuning
  - Normalize the loss based on the total number of batches across all GPUs, same as in the conformer recipe of TorchAudio.
  - Use mixed precision training.
  - Add "|" after the end of transcription to capture the silence/word termination, same as in fairseq recipe.

- Update the WER results on LibriSpeech dev and test sets.

|                   | WER% (Viterbi)|  WER% (KenLM) |
|:-----------------:|--------------:|--------------:|
| dev-clean         |       10.9    |       4.2     |
| dev-other         |       17.5    |       9.4     |
| test-clean        |       10.9    |       4.4     |
| test-other        |       17.8    |       9.5     |

Pull Request resolved: https://github.com/pytorch/audio/pull/2744

Reviewed By: carolineechen

Differential Revision: D40282322

Pulled By: nateanl

fbshipit-source-id: 4723584c912e70e8970149fe09de005385eaab90

Summary:
- The optimizer in fine-tuning recipe should also be `AdamW`. See https://github.com/pytorch/audio/pull/2412
- Fix the import of `DistributedBatchSampler` in hubert dataset
- Fix `dataset_path` in fine-tuning module.

Pull Request resolved: https://github.com/pytorch/audio/pull/2588

Reviewed By: carolineechen

Differential Revision: D38243423

Pulled By: nateanl

fbshipit-source-id: badc88ce9eddfd71270201a65ae89433fae2733f

Summary:
Meta:
**If you take no action, this diff will be automatically accepted on 2022-06-23.**
(To remove yourself from auto-accept diffs and just let them all land, add yourself to [this Butterfly rule](https://www.internalfb.com/butterfly/rule/904302247110220))

Produced by `tools/arcanist/lint/codemods/black-fbsource`.

#nocancel

Rules run:
- CodemodTransformerSimpleShell

Config Oncall: [lint](https://our.intern.facebook.com/intern/oncall3/?shortname=lint)
CodemodConfig: [CodemodConfigFBSourceBlackLinter](https://www.internalfb.com/code/www/flib/intern/codemod_service/config/fbsource_arc_f/CodemodConfigFBSourceBlackLinter.php)
ConfigType: php
Sandcastle URL: https://www.internalfb.com/intern/sandcastle/job/13510799586951394/
This diff was automatically created with CodemodService.
To learn more about CodemodService, check out the [CodemodService wiki](https://fburl.com/CodemodService).

_____

## Questions / Comments / Feedback?

**[Click here to give feedback about this diff](https://www.internalfb.com/codemod_service/feedback?sandcastle_job_id=13510799586951394).**

* Returning back to author or abandoning this diff will only cause the diff to be regenerated in the future.
* Do **NOT** post in the CodemodService Feedback group about this specific diff.

drop-conflicts

Reviewed By: adamjernst

Differential Revision: D37375235

fbshipit-source-id: 3d7eb39e5c0539a78d1412f37562dec90b0fc759

Summary:
The PR contains the CTC fine-tuning recipe of HuBERT Base model.
The files include:
- lightning module
- training script
- README and the result table
- evaluation scripts

Pull Request resolved: https://github.com/pytorch/audio/pull/2352

Reviewed By: hwangjeff

Differential Revision: D36915712

Pulled By: nateanl

fbshipit-source-id: 0249635ad5e81a8aa2d228c1d5fe84d78b62a15b