Repository navigation
Fix increasing scale groups in English number normalization - #2896
Open
Shubham-Padkonde wants to merge 1 commit into
Open
Shubham-Padkonde wants to merge 1 commit into
Shubham-Padkonde wants to merge 1 commit into
Conversation
ravindra-y
added a commit
to ravindra-y/pr-courtroom
that referenced
this pull request
Oct 24, 2026
Fix live trials on real GitHub PRs (e.g. openai/whisper#2896)
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The English number normalizer currently turns
one thousand millionand1000 millioninto1000, rather than1000000000. Its integer multiplier path always preserves the part above the last three digits, so a followingmillionorbillionnever scales the thousands that belong to it.Use the current multiplier as the group boundary for multipliers above a thousand. This also handles ordinal/plural multiplier forms and retains already accumulated higher groups:
one trillion two thousand millionbecomes1002000000000. Existing descending expressions such asone million two hundred thousandremain unchanged.The new parameterized cases run through both
EnglishNumberNormalizerandEnglishTextNormalizer. Sixteen cases fail before the fix; all 24 normalizer cases pass afterward.Validation: all 37 tests excluding transcription-model and CUDA tests pass on Python 3.10 / PyTorch 2.6 CPU. The repository's pinned pre-commit hooks pass for both changed files. Model transcription and CUDA tests were not run.
git diff --checkpasses.AI-assisted contribution.