Microsoft-backed G42, an AI technology company based in Abu Dhabi, has unveiled “Nanda,” a large language model (LLM) created especially for Hindi speakers in India. Both national and international dignitaries were present at the unveiling, which took place at the UAE-India Business Forum in Mumbai.
Nanda is a 13-billion-parameter LLM that was trained on an enormous dataset of 2.13 trillion tokens, which included Hindi, English, and Hinglish. The model is named after Nanda Devi, the second-highest peak in India.
Nanda’s release comes after G42 released Jais, an open-source Arabic language model intended for Arabic natural language processing (NLP), last year. Several model sizes, including one with 70 billion parameters, were included in the Jais release.
Microsoft became a major supporter and secured a seat on the board of G42 with a $1.5 billion investment in April 2024. Microsoft’s dedication to developing AI capabilities and growing its presence in emerging markets is demonstrated by this investment.
Also Read:
Factors to Watch Next Week: FIIs, Rate Cut, and Economic Data
The Hidden Potential of Abandoned Projects: From Failure to World-Class Triumphs
