AI4Bharat, Bodhan AI Create 1.2 Billion Parameter Speech Model

TL;DR. AI4Bharat and Bodhan AI collaborated to develop an open-source 1.2 billion parameter speech model supporting 26 Indian languages. - This model focuses on automatic speech recognition (ASR) and text-to-speech (TTS) for regional language inclusion. - It leverages a large, multi-domain dataset to enhance accuracy across diverse accents and speaking styles. - The initiative aims to democratize AI accessibility for India's linguistic diversity and reduce technological barriers.

Sources

Back to QLANKR News