BHASHINI and IISc Bangaluru : Text-to-Speech Model for Sourashtra Language

The Digital India BHASHINI Division under the Ministry of Electronics and Information Technology, in collaboration with the Indian Institute of Science, Bengaluru, has developed Text-to-Speech models for the Sourashtra language. Utilizing the VITS architecture and trained on a dedicated speech corpus of over 28 hours from the SPIRE Lab at IISc, the models feature both male and female voices available in Tamil and Sourashtra scripts. This initiative aims to support digital access, language preservation, and technological inclusion for the Sourashtra-speaking community, which had a limited digital presence, while making resources accessible via BHASHINI's AIKOSH platform.