Tshiamo19/sepedi-translation-eval

This project evaluates the translation performance of two state-of-the-art multilingual machine translation models. The target language is Sepedi (Northern Sotho). We used the Corrected FLORES dev set to benchmark translation quality. The evaluation includes both quantitative metrics and qualitative error analysis. # sepedi-translation-eval This project evaluates the translation performance of two state-of-the-art multilingual machine translation models. The target language is Sepedi (Northern Sotho). We used the Corrected FLORES dev set to benchmark translation quality. The evaluation includes both quantitative metrics and qualitative error analysis.