MIC-Off / datasets /opus /README.md
VethaNarayananG's picture
Refactor translation system to modular v3.0 architecture with offline NLLB/Whisper, gTTS fallback, corrections, and automated tests
b97b788
|
Raw
History Blame Contribute Delete
738 Bytes

A newer version of the Gradio SDK is available: 6.24.0

Upgrade

OPUS Parallel Corpus

Parallel sentence pairs from OPUS (the largest open parallel corpus).

Download

https://opus.nlpl.eu/

Expected Format

CSV Format

source_lang,target_lang,source_text,target_text
en,fr,Hello,Bonjour
en,ta,Good morning,காலை வணக்கம்

JSON Format

[
  {"source_lang": "en", "target_lang": "fr", "source_text": "Hello", "target_text": "Bonjour"},
  {"source_lang": "en", "target_lang": "ta", "source_text": "Good morning", "target_text": "காலை வணக்கம்"}
]

TSV Format (tab-separated)

en\tfr\tHello\tBonjour
en\tta\tGood morning\tகாலை வணக்கம்

Place your files in this folder and restart the application.