MahaMarathi 7B: All You Need To Know About Latest Marathi LLM

MahaMarathi 7B is a free and open base Marathi Large Language Model available on Hugging Face. After indic LLMs like Telugu, Malayalam, Tamil, and Oriya Llama, MahaMarathi 7B is in the race.

This regional language AI platform is a collaboration between Dr. Aakash Patil, a postdoctoral researcher at Stanford University, Mrunmayee Shende, cofounder of CourtEasy.ai, and Niraj Singh, ML engineer at Inbound Health, to empower Maharashtra and India with artificial intelligence. 

This article includes information about the MahaMarathi Large Language Model, which makes AI more accessible and applicable to non-English languages. 

What is MahaMarathi 7B?

MahaMarathi 7B is a domain-adapted, continually pre-trained, and instruction fine-tuned native Marathi Large Language Model (LLM). This open-source LLM is domain-adapted and continuously pre-trained, and its instructions are fine-tuned using the Meta Llama-2 and Mistral AI frameworks. Also, it is equipped with 7 billion parameters and demonstrates superior performance in various natural language processing tasks.

This AI-based Large Language Model is capable of handling complex conversations and instructions in Marathi, a language spoken by more than 83 million people. To promote broad access and applications of Marathi LLM to democratize machine learning research for businesses and e-governance, the company released the initial version of the pre-trained base model on HuggingFace.

Google Announces Gemma: All About The Latest Two LLMs

They are encouraging startups and organizations to innovate by developing fine-tuned models for various use cases, and we are happy to offer support or advice. Also, the company has planned to release the instruction-tuned and preference-optimized models in the coming months.

MahaMarathi 7B is appropriate for handling complex conversations and instructions because it takes into account the cultural context, unique linguistic features, and complexity of Marathi. The language model available for free promotes broader access and encourages applications in various fields, including business and e-governance.

Maharashtra is a powerhouse and the largest contributor to India’s economy, with Marathi businesses and consumers contributing over 15% of the Indian GDP and a rapidly growing state economy at a CAGR of over 10% with a GSDP of over ₹38.79 trillion (US$486 billion).

The potential effects of this Marathi LLM on a variety of industries, including skill development, education, healthcare, agriculture, the environment, urban planning, and traffic management, are expected to increase the existing number exponentially. 

AI is becoming more widely available and applicable to non-English languages with the release of the Marathi LLM. In the upcoming months, the team intends to release models that have been preference- and instruction-optimized for better implication of the indic LLMs.

What is Google ‘localllm’: Develop Gen AI App without GPUs