HomeTechnologySarvam AI Debuts Flagship Open-Source LLM With 24 Billion Parameters

Sarvam AI Debuts Flagship Open-Source LLM With 24 Billion Parameters

 Bengaluru  Bengaluru In a major leap in India’s AI goals, Sarvam AI has unveiled their most popular open source Large Language Model (LLM) with 24 billion of parameters. The model puts its Indian AI startup in the international spotlight and joins the likes of OpenAI, Google, and Anthropic in the race to develop the next generation of artificial intelligence (AI). AI.

India’s Biggest LLM — Open-Source and Homegrown

Sarvam AI’s LLM designed entirely in India and designed to support the multilingual comprehension and understanding and optimized performance in Indian dialects and usage scenarios. The model is open-source and can be deployed that makes it accessible to developers, researchers as well as enterprises looking to create custom AI applications without locking in to a particular vendor.

 “We built this for India and the world,” Sarvam AI’s co-founder said and co-founder, pointing out that it is trained using different datasets that include Indic scripts.

 External Sources: TechCrunch – India’s Sarvam AI enters LLM arena

Why This Matters: A Boost to India’s AI Ecosystem

The launch coincides with a period when India is pursuing the sovereignty of AI and is encouraging local businesses to develop models that are tailored to the needs of India — which include healthcare as well as governance and education. Sarvam AI’s LLM is expected to function as the basis template for a range of applications ranging from chatbots to the generation of content and tools for translation.

As per AI researchers the model drastically reduces the time to inference and costs of computation in comparison to other global models and offers a more cost-effective option with superior accuracy for low-resource situations.

Tech Highlights

  •   24B parameter that are trained on domain-specific and multilingual datasets

  •   Optimized to allow for the GPU’s efficient computation inference and for deployment using cloud infrastructures that are standard

  •   It is compatible with Transformers and the PyTorch ecosystem

  •   Released under an an open-source licence which allows for the community to drive development

 External Sources: Analytics India Magazine Sarvam AI’s model 24B creates new benchmarks

A Rising Force in Global AI

With this announcement, Sarvam AI joins the ranks of emerging AI companies that are based in across the Global South. Sarvam AI has also hinted at additional releases in the parameter 70B+ range as well as collaborations with universities and tech companies to increase the adoption of AI.

RELATED ARTICLES
- Advertisment -

Most Popular

Recent Comments