Logo Lanfrica

aisod/AISOD-AI-LRLM

Domain:

natural language processing

Record type:

model
Creator:
ais
Host:
This is the AISOD AI Low Resource and Reasoning Language Model thats fed by NAMQULA Language Game, contributors and community. # AISOD AI-LRLM ## AI Language Model for Low-Resource Languages (Namibian Languages) **Model Name:** AISOD AI-LRLM **Project Goal:** Build an offline/online AI Language Model for Namibian Languages (starting with Oshiwambo) **Target Application:** Namqula Language Learning App --- ## Quick Start ### 1. System Requirements - Python 3.10 or higher - 8GB+ RAM (16GB recommended) - 20GB+ free disk space - Core i5 8th gen or equivalent (CPU training supported) ### 2. Initial Setup ```bash # Check your system python scripts/check_system.py # Install dependencies pip install -r requirements.txt # Or use the setup script (Windows) python setup.py ``` ### 3. Process Your Data ```bash # Process the Oshiwambo data file python scripts/process_data.py --input "Oshiwambo Data 1.txt" --output data/processed/ ``` ### 4. Download Base Model ```bash # Download Llama 3.2 3B (recommended to start) python scripts/download_model.py --model llama-3.2-3b ``` ### 5. Start Training ```bash # Train with your data (CPU-optimized) python training/train.py --config config/llama-3.2-3b.yaml --data data/processed/ ``` --- ## Project Structure ``` AISOD-AI-LRLM/ ├── README.md # This file ├── requirements.txt # Python dependencies ├── setup.py # Setup script ├── .env.example # Environment template ├── .gitignore # Git ignore file │ ├── config/ # Configuration files │ ├── models.yaml # Model registry │ ├── llama-3.2-3b.yaml # Llama config (CPU-optimized) │ └── mistral-7b.yaml # Mistral config │ ├── data/ # Data storage │ ├── raw/ # Original data files │ ├── processed/ # Processed training data │ └── README.md # Data format guide │ ├── models/ # Model storage │ ├── base/ # Base models (downloaded) │ ├── fine-tuned/ # Fine-tu …