The Indus Project | Tech Mahindra

Why Indus Project

Indus is a civilizational initiative tailored for India to empower all Indic languages originating from The Indus Valley Civilization. The project's dual objectives are to construct a language model deeply rooted in Indian culture and, secondly, to excel in prevailing benchmarks. It is our endeavor to create a foundation model for Indian languages that model that will enable and simplify communication across the country and preserve our languages and dialects.More

Indus is a civilizational initiative tailored for India to empower all Indic languages originating from The Indus Valley Civilization. The project's dual objectives are to construct a language model deeply rooted in Indian culture and, secondly, to excel in prevailing benchmarks. It is our endeavor to create a foundation model for Indian languages that model that will enable and simplify communication across the country and preserve our languages and dialects. In the first phase, we will create the LLM for Hindi and 37+ dialects and then move ahead in a phased manner to cover other languages and dialects.

Less

India’s Language Graph at a Glance

0
Officially Recognized Languages
0 +
Dialects Spoken in India
0
Unofficial Dialects

Potential Benefits

Farmer’s Network

A digital buddy for over 140 million farmers that will provide the required information on loans, pesticides, and more.

JAM Stack Connect for India

A solution stack architecture designed to make the web faster, easier to scale, and more secure.

Education Enablement

Where we aid children and enable them to understand various related subjects better.

Industry Foundation Models

Verticalized foundation models are created for specific industries like media, telecom, and healthcare.

Dialect Preservation

We help preserve dialects that are spoken by many but are not digitized.

Rural Finance

A rural kiosk can decipher speech in local dialects to solve different financing problems. In the absence of India’s foundation model, this would increase the OPEX for many companies.

Mobile Conversational Systems

Mobile conversational systems will be embedded in different equipment to make it more interactive and conversational in local dialects.

Public Healthcare Infrastructure

Ensuring ethical and useful information is made available transparently in the local dialect.

What We Offer

Open-Source Model for India and the World

The model has been launched for beta testing within Tech Mahindra with 539 million parameters and 10 billion pure Hindi + dialect tokens. Our model will be launched for open source in February 2024.

Text Generation to Chat Model Integration

The model in the first phase will be a decoder only to generate text, and the next phase will include RLHF to convert it into a chat model.

Voice Capability Enablement

The last phase will provide voice to the model.

Making the Headlines

Induproject
Tech Mahindra Aims to Create Foundational Language Model Rooted in India

Project Indus aims to create foundational language model focused on Indian languages following an open-source approach.

openai-rival
Tech Mahindra to Launch OpenAI Rival ‘Project Indus’ Early Next Year

The 15-member Project Indus team has gathered 1.2 terabytes of data in Hindi and related dialects.

Induproject
Tech Mahindra to Launch a "Made In India" ChatGPT Model

Indus is a civilizational initiative tailored for India to empower all Indic languages that have originated out of The Indus Valley Civilization.

Induproject
Tech Mahindra Set to Launch Indus, Its Indigenous Rival Of ChatGPT

The primary objective is to establish a 7-billion parameter linguistic model to make advancements in NLP.

Get In Touch

Need more information?
We will take approximately 3-5 working days to respond to your enquiry.

By clicking on the submit button, you agree with the privacy policy.