All about BharatGPT’s ‘Hanooman’ in brief:
- The BharatGPT group, led by IIT Bombay and seven other Indian engineering institutes, is launching Hanooman, a ChatGPT-like service supported by Reliance Industries Ltd and the Department of Science and Technology.
- Hanooman is a series of large language models (LLMs) capable of responding in 11 Indian languages initially, expanding to over 20. It operates in fields like healthcare, governance, financial services, and education.
- It's a multimodal AI tool generating text, speech, videos, etc., in multiple Indian languages. One variant, VizzhyGPT, is tailored for healthcare using medical data.
- Model sizes range from 1.5 billion to 40 billion parameters.
- Challenges include the quality of datasets in Indian languages, often synthetic and prone to inaccuracies.
- Other Indian language model projects exist, like Sarvam and Krutrim.
- Large language models (LLMs) use deep learning to process extensive text data, learning meanings and relationships between words from training datasets like Wikipedia and OpenWebText.
Thanks for reading Hanooman.
