Skip to content

CommonVoice FR - Preprocessing stripped every lowercase character in the model #2085

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Kizyow wants to merge 1 commit into
base: master
Choose a base branch
from
+3 −2
Open

CommonVoice FR - Preprocessing stripped every lowercase character in the model #2085

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions egs/commonvoice/ASR/local/preprocess_commonvoice.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,7 +49,8 @@ def normalize_text(utt: str, language: str) -> str:
if language == "en":
return re.sub(r"[^a-zA-Z\s]", "", utt).upper()
elif language == "fr":
return re.sub(r"[^A-ZÀÂÆÇÉÈÊËÎÏÔŒÙÛÜ' ]", "", utt).upper()
utt = utt.upper()
return re.sub(r"[^A-ZÀÂÆÇÉÈÊËÎÏÔŒÙÛÜ' ]", "", utt)

@gemini-code-assist gemini-code-assist Bot Jun 2, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

medium

The French uppercase letter Ÿ (corresponding to lowercase ÿ) is missing from the allowed characters list in the regular expression. Although rare, it is used in French proper nouns (e.g., L'Haÿ-les-Roses, Moÿ-de-l'Aisne). Without it, any occurrence of ÿ or Ÿ will be silently stripped after conversion to uppercase.

Adding Ÿ to the character class ensures complete coverage of French alphabet diacritics.

Suggested change
return re.sub(r"[^A-ZÀÂÆÇÉÈÊËÎÏÔŒÙÛÜ' ]", "", utt)
return re.sub(r"[^A-ZÀÂÆÇÉÈÊËÎÏÔŒÙÛÜŸ' ]", "", utt)

elif language == "pl":
return re.sub(r"[^a-ząćęłńóśźżA-ZĄĆĘŁŃÓŚŹŻ' ]", "", utt).upper()
elif language in ["yue", "zh-HK"]:
Expand Down Expand Up @@ -139,7 +140,7 @@ def preprocess_commonvoice(
if partition == "validated":
logging.warning(
"""
The 'validated' partition contains the data of both 'train', 'dev'
The 'validated' partition contains the data of both 'train', 'dev'
and 'test' partitions. We filter out the 'dev' and 'test' partition
here.
"""
Expand Down
Loading