The model is trained to answer questions about health topics (Open-book question answering-comprehend). cointegrated/rut5-base-multitask For training, a compact T5 model was used: cointegrated/rut5-base-multitask

The training was conducted on a small set out of 220 thousand pairs of question-answer sentences, so it still does not work as correctly as we would like.

The model is not a medical application and it is strongly discouraged to use the model for medical purposes!

@misc{UARITM, title={sentence-transformers: Semantic similarity of medical texts ukr, kor, eng}, author={Uaritm}, year={2025}, url={https://ai.esemi.org}, }