Orama Fine Tuned Models

This is a collection of fine-tuned models used by Orama.

Getting started

The models in this repository have been trained with python3.12. Therefore, we recommend using a venv with that specific Python version.

python3.12 -m venv .venv
source .venv/bin/activate

Then install all the dependencies:

pip install -r requirements.txt

You can generate synthetic data with the synthetic.py file always included with every model. It depends on either a local or remote Ollama installation for generating synthetic data via Qwen2.5 14B.

Example:

cd models/query-translator
OLLAMA_URL=http://localhost:11434 python synthetic.py

Models

• Query Translator Mini: HuggingFace

License

AGPLv3