Airavata: Introducing Hindi Instruction-tuned LLM

TL;DR

Airavata is an instruction-tuned large language model for Hindi, created through fine-tuning on diverse datasets, with accompanying benchmarks and datasets to advance research in Indic languages.

Contribution

The paper introduces Airavata, a novel Hindi instruction-tuned LLM, along with the IndicInstruct dataset and evaluation framework, facilitating further research in Indic language models.

Findings

Airavata performs well on Hindi assistive tasks.

IndicInstruct dataset enables diverse instruction tuning.

Framework allows comprehensive evaluation of Hindi LLMs.

Abstract

We announce the initial release of "Airavata," an instruction-tuned LLM for Hindi. Airavata was created by fine-tuning OpenHathi with diverse, instruction-tuning Hindi datasets to make it better suited for assistive tasks. Along with the model, we also share the IndicInstruct dataset, which is a collection of diverse instruction-tuning datasets to enable further research for Indic LLMs. Additionally, we present evaluation benchmarks and a framework for assessing LLM performance across tasks in Hindi. Currently, Airavata supports Hindi, but we plan to expand this to all 22 scheduled Indic languages. You can access all artifacts at https://ai4bharat.github.io/airavata.