NVIDIA Launches Nemotron-4-Mini-Hindi-4B Model for AI in India

NVIDIA has launched the Nemotron-4-Mini-Hindi-4B model, a small language model for Hindi, enabling businesses to deploy AI solutions specific to local needs. This model, part of NVIDIA’s NIM microservice, can be deployed on NVIDIA GPU-accelerated systems, optimising performance for various applications. 

Tech Mahindra is the first to implement this model, creating Indus 2.0, which focuses on Hindi and its dialects. 

The Nemotron Hindi model has 4 billion parameters and was derived from a 15-billion parameter multilingual model, Nemotron-4. It was trained with real-world Hindi data and synthetic data, including English. 

After fine-tuning with NVIDIA NeMo, it leads on multiple accuracy benchmarks for AI models with up to 8 billion parameters. Packaged as a microservice, it supports various industry applications, including education and healthcare.

Innovators and enterprises in India are utilising NVIDIA NeMo to create customised language models. 

Sarvam AI has developed Sarvam 1, India’s first multilingual language model, trained on NVIDIA H100 GPUs. This model supports English and ten major Indian languages. 

Gnani.ai has built a multilingual speech-to-speech model that serves as an AI customer service assistant, handling approximately 10 million real-time interactions daily.

Large enterprises are also using NeMo. Flipkart integrates NeMo Guardrails into its conversational AI systems for enhanced safety. Krutrim is developing a multilingual foundation model using Mistral NeMo 12B. Zoho Corporation plans to use NVIDIA TensorRT-LLM and Triton Inference Server for its language models.

Meanwhile, Tata Consultancy Services and Wipro are offering NVIDIA NeMo-accelerated solutions across various industries. TCS is creating domain-specific language models for telecommunications, retail, and financial services. Wipro is developing custom conversational AI solutions for customer service interactions.