International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 186 - Number 62 |
Year of Publication: 2025 |
Authors: Olawale Timothy Adeboje, Olawale Timothy Adeboje, Olusola Adebayo Adetunmbi, Raphael Olufemi Akinyede |
10.5120/ijca2025924463 |
Olawale Timothy Adeboje, Olawale Timothy Adeboje, Olusola Adebayo Adetunmbi, Raphael Olufemi Akinyede . A Deep Learning-based English to Yoruba Neural Translation Model. International Journal of Computer Applications. 186, 62 ( Jan 2025), 38-41. DOI=10.5120/ijca2025924463
Yoruba is one of the most widely spoken languages in Africa and certain parts of the world. However, the dominance of the English language in Nigeria has contributed to the gradual decline of Yoruba language. This research was motivated by the need to address this by developing a Yoruba neural translation that aim to translate text written in English Language to Yoruba language. The research utilized a parallel corpus obtained from MENYO-20k and odunola/yoruba-english pairs (https://huggingface.co/datasets/odunola/yoruba-englishpairs). Transformer model was used, Bidirectional Encoder Representations from Transformers (BERT) was used for the encoder and T5-Base model on the decoder side of the transformer. The developed model achieved the BLUE score of 72%, which means a strong alignment between the translated outputs and reference texts, reflecting the model’s capability to maintain the integrity of the original message.