Skip to content

This project is about a neural machine translation in wolof language to french .

License

Notifications You must be signed in to change notification settings

abdouaziz/wolof-translation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Wolof Translation

Description

Machine translation is an area of NLP research that allows one language to be translated into another language. Today it is difficult to find an automatic translator for languages that are not well represented, especially in Africa. This is the purpose of this project.

Wolof is a language spoken in Senegal and many other countries in West Africa. However, having a translation tool would make it possible to popularise works written in Wolof, but also to contribute to the representation of the language's resources.

Objective

To achieve our objects in the project we will implement two models based on the architecture of transformers, namely :

  • A model based on seq2seq transformers for more information I invite you to visit the site

  • A model based on T5

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

install the dependencies for this project by running the following commands in your terminal:

 pip install -r requirements.txt

run the seq2seq model by running the following command in your terminal:

python src/seq2seq.py --train_file="./input/wolof.csv" \
                        --max_source_length=150 \
                        --max_target_length=150 \
                        --number_epochs=4 \
                        --learning_rate=3e-8 \
                        --epsilone=1e-9 \
                        --train_batch_size=3 \
                        --model_name="bert-base-cased"

run the t5 model by running the following command in your terminal:

python src/t5.py --train_file="./input/wolof.csv" \
                        --max_source_length=150 \
                        --max_target_length=150 \
                        --number_epochs=4 \
                        --learning_rate=3e-8 \
                        --epsilone=1e-9 \
                        --train_batch_size=3 \
                        --model_name="t5-small" \
                        --task_prefix="translation French to Wolof: "

Ressources

Attention Is All You Need

Transformers-based Encoder-Decoder Models

A Review of the Neural History of Natural Language Processing

Neurail Machine Translation By Jointly Learning to Align and Translate

Sequence to Sequence Learning with Neural Networks

About

This project is about a neural machine translation in wolof language to french .

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages