Opus Mt Rw En
A Transformer-based neural machine translation model for Kinyarwanda (rw) to English (en), trained by the Helsinki-NLP team using the OPUS dataset.
Downloads 172
Release Time : 3/2/2022
Model Overview
This model is specifically designed for translating Kinyarwanda text into English, employing the standard Transformer-align architecture and supporting sentence-level translation tasks.
Model Features
Alignment Attention Mechanism
Uses the transformer-align architecture to generate word alignment information between source and target languages.
Standardized Preprocessing
Employs SentencePiece for text standardization to enhance translation quality.
Multi-domain Coverage
Training data includes JW300 religious texts and Tatoeba everyday phrases, covering diverse domains.
Model Capabilities
Kinyarwanda to English text translation
Sentence-level translation
Cross-language information conversion
Use Cases
Religious Text Translation
JW300 Bible Content Translation
Translates religious literature from Kinyarwanda to English
BLEU 37.3 (JW300 test set)
Daily Communication Translation
Tatoeba Sentence Translation
Translation of daily conversations and simple sentences
BLEU 49.8 (Tatoeba test set)
Featured Recommended AI Models
Š 2025AIbase