Question 1

Does Maha-1 use fine-tuning at all?

Accepted Answer

Fine-tuning may be used for language, tone and reasoning style, particularly for India's languages, but factual grounding relies on retrieval, not fine-tuned knowledge.

Question 2

Is RAG slower than a fine-tuned model?

Accepted Answer

RAG adds a retrieval step, which adds some latency, but this is a small trade-off for grounded, sourced, auditable answers, essential for government use.

Question 3

Can a fine-tuned model be updated as fast as a RAG system?

Accepted Answer

No. Updating a RAG system's knowledge base is close to immediate. Fine-tuning requires retraining, which is slower and more resource-intensive.