Lexaze/ET-Amharic-Multilingual-Language-Model-for-Public-Services

# 🇪🇹 Amharic Multilingual Language Model for Public Services A demo-ready portfolio project showing how an Amharic-first multilingual AI assistant can help citizens access public-service information. > **Disclaimer:** This is a demo project, not an official government system. It uses a small local sample knowledge base and should not be used as legal, immigration, tax, health, or government advice. ## Features - Amharic / English public-service chatbot - Retrieval-Augmented Generation using a local JSON knowledge base - Multilingual semantic search with `sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2` - FAISS vector search - FastAPI backend - Streamlit web UI - Retrieved source display - Hallucination safeguards through grounded response composition - Tests with Pytest - Docker support ## Supported demo service topics - ID renewal - Passport application - Tax registration - Business license - Health center information - Education document certification - Social services referral ## Project structure ```text amharic-public-services-llm/ ├── README.md ├── requirements.txt ├── .env.example ├── app/ │ ├── __init__.py │ ├── main.py │ ├── rag.py │ ├── translator.py │ ├── prompts.py │ └── utils.py ├── data/ │ └── public_services_knowledge_base.json ├── frontend/ │ └── streamlit_app.py ├── tests/ │ └── test_api.py ├── docs/ │ ├── architecture.md │ └── demo_script.md └── Dockerfile ``` ## Quick start ### 1. Clone and enter the project ```bash git clone cd amharic-public-services-llm ``` ### 2. Create a virtual environment ```bash python3 -m venv .venv source .venv/bin/activate ``` On Windows PowerShell: ```powershell python -m venv .venv .venv\Scripts\Activate.ps1 ``` ### 3. Install dependencies ```bash pip install --upgrade pip pip install -r requirements.txt ``` The first run downloads the multilingual embedding model from Hugging Face. ### 4. Configure environment ```bash cp .env.example .env ``` ### 5. Run the …