Authors: Bogdan Tewunalany, Ali Berberov
Github: https://github.com/TBSj/Qarachay_Malqar_translator_python

As a base we took mbart-50 model and trained it on 27235 parallel sentences from russian to Qarachay-Malqar language. It is not enough for good prediction, nevertheless, there is some result.

Nowdays, we are collecting more sentences to improve our result. And, I hope, it gives us opportunity to not use pre-trained models and realize translator on R instead of python.

Dataset: https://huggingface.co/datasets/TSjB/qm_ru_parallel

Where to use:

https://huggingface.co/spaces/TSjB/QM_RU_translator

https://tsjb-qm-ru-translator.hf.space/

Telegram: https://t.me/QMKochBot