[Whisper] Move decoder id method to tokenizer#20589
Merged
sanchit-gandhi merged 1 commit intohuggingface:mainfrom Dec 5, 2022
Merged
[Whisper] Move decoder id method to tokenizer#20589sanchit-gandhi merged 1 commit intohuggingface:mainfrom
sanchit-gandhi merged 1 commit intohuggingface:mainfrom
Conversation
sanchit-gandhi
commented
Dec 5, 2022
Comment on lines
+402
to
+403
| elif self.language in TO_LANGUAGE_CODE.values(): | ||
| language_id = self.language |
Contributor
Author
There was a problem hiding this comment.
Processor's get_decodoer_prompt_ids expected a language code id (e.g. "es"). Tokenizer's set_prefix_tokens expected a language (e.g. "Spanish"). This PR amends the tokenizer method to handle either.
ArthurZucker
approved these changes
Dec 5, 2022
Collaborator
ArthurZucker
left a comment
There was a problem hiding this comment.
Very nice thanks a lot! I remember it was temporary so that's a nice follow up!
|
The documentation is not available anymore as the PR was closed or merged. |
sgugger
approved these changes
Dec 5, 2022
Collaborator
sgugger
left a comment
There was a problem hiding this comment.
Nice fix and thanks for making it fully backward compatible!
5 tasks
This was referenced Dec 5, 2022
mpierrau
pushed a commit
to mpierrau/transformers
that referenced
this pull request
Dec 15, 2022
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Moves the method
get_decoder_prompt_idsfrom the processor to the tokenizer. The primary reason for this change is that the ASR pipeline class does not load the processor object, but rather the feature extractor and tokenizer separately (see docs). Therefore, as things currently stand, pipeline does not have access to the processor methodget_decoder_prompt_ids. By moving it to the tokenizer, it will be able to call this method with pipeline.Note that this is not a breaking change: we retain a method
get_decoder_prompt_idsin the processor. This method simply calls theget_decoder_prompt_idsfrom the tokenizer:transformers/src/transformers/models/whisper/processing_whisper.py
Lines 44 to 45 in ca8b332
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.