The researcher of the Language Technologies Unit, Martí Llopart, presents the Matxa speech synthesis model, published in April this year, at the conference “Data and AI Proof of Concept” organized by CIDAI.
The Matxa model is the first open-source speech solution that includes the different dialectal variants of Catalan. It is a model trained on MareNostrum 5 and represents a step forward in the presence of the plurality of Catalan in artificial intelligence.
