English translation to Igbo using TensorFlow Transformer model.
This NLP project applies the Tensorflow Transformer model in a sequence-to-sequence modelling task. Sports headlines published in English language are translated into Igbo language using the transformer model and the results are compared to the published Igbo version. Igbo is the native language spoken by people from the south eastern part of Nigeria. It is spoken by over 31 million people.
Project was created with:
- Python==3.8.18
- TensorFlow==2.13.0
- Keras==2.13.1
The data, made up of headlines of sports news, was scrapped from the Complete Sports English and Igbo websites. The 17963 English-Igbo pairs was split into training(80%), validation(10%) and test(10%) datasets.