Towards Comprehensive Vietnamese Retrieval-Augmented Generation and Large Language Models

TL;DR

This paper advances Vietnamese language understanding and generation by developing open datasets and pre-trained models for Retrieval-Augmented Generation and Large Language Models.

Contribution

It introduces new open datasets and pre-trained models specifically designed for Vietnamese RAG and LLMs, enhancing language technology resources.

Findings

Improved Vietnamese language understanding capabilities

Enhanced generation quality for Vietnamese texts

Availability of open datasets and models for research

Abstract

This paper presents our contributions towards advancing the state of Vietnamese language understanding and generation through the development and dissemination of open datasets and pre-trained models for Vietnamese Retrieval-Augmented Generation (RAG) and Large Language Models (LLMs).