Yaduha translation tutorial

An interactive Jupyter notebook that walks through the LLM-Assisted Rule-Based Machine Translation (LLM-RBMT) approach we used to build an English → Owens Valley Paiute translator.

Setup

Requires uv. From this directory:

uv sync                           # install yaduha, yaduha-ovp, jupyter, etc.
uv run jupyter lab tutorial.ipynb # or: uv run jupyter notebook

If you use VS Code, open the folder, open tutorial.ipynb, and pick the .venv kernel that uv sync created.

API key

Several cells call the OpenAI API. To run them, drop a .env file next to the notebook:

OPENAI_API_KEY=sk-...