Hugging Face's logo

language: - yo - en datasets: - JW300 + Menyo-20k

mbart50-large-yor-eng-mt

Model description

mbart50-large-yor-eng-mt is a machine translation model from Yorùbá language to English language based on a fine-tuned facebook/mbart-large-50 model. It establishes a strong baseline for automatically translating texts from Yorùbá to English.

Specifically, this model is a mbart-large-50 model that was fine-tuned on JW300 Yorùbá corpus and Menyo-20k. The model was trained using Swahili(sw_KE) as the language since the pre-trained model does not initially support Yorùbá. Thus, you need to use the sw_KE for language code when evaluating the model.

Limitations and bias

This model is limited by its training dataset. This may not generalize well for all use cases in different domains.

Training data

This model was fine-tuned on on JW300 corpus and Menyo-20k dataset

Training procedure

This model was trained on NVIDIA V100 GPU

Eval results on Test set (BLEU score)

Fine-tuning mbart50-large achieves 15.88 BLEU on Menyo-20k test set while mt5-base achieves 15.57

BibTeX entry and citation info

By David Adelani