Aller au contenu
login
arrow_backRetour aux issues
sohamsinha2018/ramblefix #13

Language route: English + Tagalog

ecoDébutant help wanted language route corpus needed model evaluation

descriptionDescription

Goal Explore a local English + Tagalog route for bilingual dictation. Why this matters English + Tagalog is a common code-switching pattern, and it is a good next test for whether RambleFix can generalize beyond Hindi + English. Useful contributions - Representative English + Tagalog work dictation clips and gold transcripts. - Notes on common code-switching patterns and terms. - Candidate local ASR models/runtimes that preserve both languages. - Same-WAV comparison against the current RambleFix English path. Acceptance bar - Pure English stays fast and accurate. - English + Tagalog keeps more meaning than the default English route. - No cloud dependency in the product path. - Any PR includes the eval surface and sample size. Start here Fork the repo, add a short experiment note or corpus fixture, and open a PR for review.
codeOuvre sur GitHub