Skip to content

Refactor SpeechT5 output tracing to standardized output capture - #44129

Closed
preetam1407 wants to merge 1 commit into
huggingface:mainfrom
preetam1407:preetam/speecht5-output-capturing
Closed

preetam1407 wants to merge 1 commit into
huggingface:mainfrom
preetam1407:preetam/speecht5-output-capturing

Conversation

@preetam1407

Copy link
Copy Markdown
Contributor

What does this PR do?

Refactors SpeechT5 to the standardized output tracing interface.

  • Adds @capture_outputs to base encoder/decoder forwards.
  • Adds _can_record_outputs mappings for hidden states and attentions.
  • Adds @can_return_tuple to wrapper/task forwards.
  • Removes manual output_attentions / output_hidden_states / return_dict plumbing.

Fixes #43979

Who can review?

@eustlb @ebezzam @vasqu @molbap

@github-actions

Copy link
Copy Markdown
Contributor

[For maintainers] Suggested jobs to run (before merge)

run-slow: speecht5

@ArthurZucker

Copy link
Copy Markdown
Collaborator

Thanks for the SpeechT5 refactor, @preetam1407! The #43979 call to contributions was closed right after this was opened and we couldn't review all the PRs it produced, so I'm closing this one. Sorry it didn't get looked at in time.

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Call to contributions: refactor output tracing in transformers

2 participants