Logo Lanfrica

igttsPro/igede_translate_project

Domain:

natural language processing

Record type:

project
Creator:
igt
Host:
This project aims to build a machine translation system and digital dataset for the **Igede language** spoken in Benue State, Nigeria. It's designed to support language preservation, data accessibility, and AI/ML-based translation. # πŸ—£οΈ Igede Language Translation Project This project aims to build a machine translation system and digital dataset for the **Igede language** spoken in Benue State, Nigeria. It's designed to support language preservation, data accessibility, and AI/ML-based translation. ## πŸ“‚ Project Structure - `scripts/`: Web scrapers and data cleaning tools. - `data/raw/`: Unprocessed files from Glosbe, Bible, dictionaries. - `data/cleaned/`: Sentence-aligned translation pairs. - `notebooks/`: Exploration, cleaning, and visualizations. - `models/`: Any trained model artifacts. - `app/`: Frontend/backend code (e.g., Streamlit demo). ## Project Structure: igede_translate_project/ β”‚ β”œβ”€β”€ data/ β”‚ β”œβ”€β”€ raw/ # Unprocessed data (scraped text, raw Bible texts, dictionaries) β”‚ └── cleaned/ # Cleaned, normalized, aligned sentence pairs for training β”‚ β”œβ”€β”€ notebooks/ # Jupyter/Colab notebooks for prototyping and EDA β”‚ └── exploratory.ipynb β”‚ β”œβ”€β”€ scripts/ # Python scripts for automation (scraping, cleaning, aligning) β”‚ β”œβ”€β”€ scrape_glosbe.py β”‚ β”œβ”€β”€ align_bible.py β”‚ └── clean_data.py β”‚ β”œβ”€β”€ models/ # NMT model files, checkpoints, tokenizer config β”‚ └── igede_en_model/ β”‚ β”œβ”€β”€ app/ # App backend or frontend (e.g., Flask, Streamlit) β”‚ └── translator_app.py β”‚ β”œβ”€β”€ requirements.txt # List of Python dependencies (requests, pandas, etc.) β”œβ”€β”€ README.md # Project overview, goals, usage instructions └── .gitignore # To ignore cache, virtualenvs, or large files ## βœ… Getting Started 1. Install dependencies: ```bash pip install -r requirements.txt