Meet Hanooman – AI Model Developed by Reliance Industries-backed BharatGPT group set to launch in March

Meet Hanooman – an AI tool developed by the BharatGPT group. It is developed with the support of Reliance Industries and top engineering schools. Its service will be available in 11 local languages across healthcare, finance, governance, and educational sectors.

What is Hanooman?

BharatGPT group backed by Mukesh Ambani’s reliance industries is set to launch its first ChatGPT style service next month, it marks a significant step for India toward the emergence of artificial intelligence.

What is Hanooman?
What is Hanooman?

Reliance Jio Infocomm and the central government developed this tool in collaboration with 8 Indian institutes of technology. A sneak peek of this model was showcased at the recent tech conference in Mumbai.

In that video, a motorcycle mechanic interacting with this model (dubbed Hanooman) in Tamil, a banker interacting with this tool in Hindi, and a Hyderabad-based software developer using this tool to write codes.

Features of Hanooman

Hanooman will operate in 11 local languages in four fields, they are,

  • Education,
  • Finance,
  • Healthcare,
  • Governance

This is not an ordinary chatbot, it is a multimodal AI tool that can generate text, speech, video, and many more in multiple Indian languages according to its developing team BharatGPT. VizzhyGPT is one of the first customized versions, it is finetuned for healthcare industries based on medical data.

The size of the AI models will range from 1.5 billion to 40 billion parameters.

Ganesh Ramakrishnan, chair of IIT Bombay’s Department of Computer Science and Engineering said, This tool will also offer text-to-speech capabilities to make it more user-friendly.

Vishnu Vardhan, the Founder of SML, during the launch of this tool, highlighted some of the issues faced prevalence of synthetic datasets. He noted the challenges in the quality of datasets in Indian languages. These synthetic datasets generated by AI instead of produced by real-time events will cause inaccuracies in data collection and analysis.

Startups like Sarvam and Krutrim are also working on open-sourced AI models for the country. These startups are backed by prominent VC investors like Lightspeed Venture Partners and billionaire Vinod Khosla’s fund.

BharatGPT’s work is somewhat unique and appreciable because it is the country’s first private-public partnership and involves major players in this field.

Jio Brain

It is reported that Reliance Jio wants to make custom AI models for various industries. Their retail telecom company also started to work Platform named “Jio Brain”, which uses AI across a network of about 450 million subscribers.

Also, read our other article about AI data platform – Alaya AI

Large Language Models or LLMs

By harnessing the power of AI, Large Language Models or LLMs generate natural-sounding responses that are learned from the vast quantity of data inputs. It is a newer type of AI popularized by the success of Openai’s ChatGPT.

These models are based on deep learning techniques to process large amounts of data. LLMs are trained to identify the meanings and relationships between the words. The greater the amount of data is fed, the more smarter and efficient the model will function.

The training data is usually large datasets like Wikipedia, Openwebtexts, and the Common Crawl Corpus. These datasets contain large amounts of text which is used by these AI models to produce natural and efficient responses.