SML & BharatGPT Ecosystem Unveils 'Hanooman': Indic AI Series at NASSCOM Event - India's First Multilingual Multimodal Large Language Model
Feb 21, 2024
SRV Media
Bombay (Maharashtra) [India], February 21: Seetha Mahalaxmi Healthcare (SML), in collaboration with IIT Bombay's BharatGPT Ecosystem, has introduced "Hanooman" (Built In India for the World) a revolutionary series of Indic AI models with multimodal capabilities. This cutting-edge technology was revealed at the prestigious NASSCOM Technology and Leadership Forum held in Mumbai.
Vishnu Vardhan, the Founder of SML, announced that the "Hanooman" series comprises Indic Language Models (LLMs) trained on 22 Indian languages, boasting sizes ranging up to 40 billion parameters. Vardhan also highlighted plans for the open-source release of these modules scheduled for the upcoming month.
Presently, "Hanooman" can proficiently interact in 11 Indian languages, including Hindi, Tamil, Telugu, Malayalam, and Marathi, with plans for expansion to encompass all 22 languages. Beyond its linguistic prowess, Hanooman stands out with its multimodal AI capabilities, enabling seamless generation of text-to-text, text-to-speech, text-to-video, and vice versa content.
Vardhan emphasized ongoing discussions with enterprises, healthcare institutions, and mobile app providers to tailor specialized models through fine-tuning, leveraging Hanooman's versatility.
However, Vardhan acknowledged the challenges posed by the quality of datasets in Indian languages, stressing the prevalence of synthetic datasets derived from translations, which may lead to inaccuracies or distortions.
The BharatGPT Ecosystem, spearheaded by IIT Bombay and supported by other IITs, along with backing from the Department of Science and Technology, SML, and Reliance Jio, underscores the collaborative effort driving the advancement of Indic AI technology.
(ADVERTORIAL DISCLAIMER: The above press release has been provided by SRV Media. ANI will not be responsible in any way for the content of the same)
