Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Whisper fine-tuning recipes for CHiME-4 and WSJ#5342

Merged
sw005320 merged 6 commits intoespnet:masterfrom
YoshikiMas:hackathon-2023s
Jul 23, 2023
Merged

Whisper fine-tuning recipes for CHiME-4 and WSJ#5342
sw005320 merged 6 commits intoespnet:masterfrom
YoshikiMas:hackathon-2023s

Conversation

@YoshikiMas
Copy link
Contributor

This PR adds configurations of Whisper fine-tuning for CHiME-4 and WSJ. I'll upload the pre-trained model later.

@kan-bayashi kan-bayashi marked this pull request as draft July 22, 2023 07:40
@YoshikiMas YoshikiMas changed the title [WIP] Whisper fine-tuning recipes for CHiME-4 and WSJ Whisper fine-tuning recipes for CHiME-4 and WSJ Jul 22, 2023
@sw005320 sw005320 added the ASR Automatic speech recogntion label Jul 22, 2023
@sw005320 sw005320 added this to the v.202307 milestone Jul 22, 2023
Copy link
Contributor

@sw005320 sw005320 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.
Please clarify whether you use an LM or not in REAMD.md

|decode_asr_lm_lm_train_lm_transformer_en_char_valid.loss.ave_asr_model_valid.acc.ave/et05_simu_beamformit_5mics|1320|126812|94.4|2.8|2.8|1.5|7.2|66.1|


## Whisper [medium_finetuning](conf/tuning/train_asr_whisper_full_warmup1500.yaml)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it with an LM?
Please remark on it.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If this one is not used, you can remove it from this PR.


## Whisper [medium_finetuning](conf/tuning/train_asr_whisper_full_warmup3000.yaml)

## Environments
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ditto
Please clarify that this does not use an LM.

@sw005320 sw005320 marked this pull request as ready for review July 22, 2023 23:02
@YoshikiMas
Copy link
Contributor Author

To clarify the point that the performance is with --cleaner whisper_en, I added a comment in each README.md. The use of the cleaner is consistent with previous whisper fine-tuning recipe.

@sw005320 sw005320 added the auto-merge Enable auto-merge label Jul 23, 2023
@codecov
Copy link

codecov bot commented Jul 23, 2023

Codecov Report

Merging #5342 (53b1f4b) into master (79a74bc) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #5342   +/-   ##
=======================================
  Coverage   76.11%   76.11%           
=======================================
  Files         672      672           
  Lines       59859    59859           
=======================================
  Hits        45563    45563           
  Misses      14296    14296           
Flag Coverage Δ
test_configuration_espnet2 ∅ <ø> (∅)
test_integration_espnet1 65.93% <ø> (ø)
test_integration_espnet2 47.92% <ø> (ø)
test_python_espnet1 20.51% <ø> (ø)
test_python_espnet2 51.35% <ø> (ø)
test_utils 23.17% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

@sw005320 sw005320 merged commit 42f3182 into espnet:master Jul 23, 2023
@sw005320
Copy link
Contributor

Thanks, @YoshikiMas!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ASR Automatic speech recogntion auto-merge Enable auto-merge ESPnet2 README

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants