Efficient In-Domain Question Answering for Resource-Constrained Environments

TL;DR

This paper introduces CRAFT, a resource-efficient retrieval augmented fine tuning method combining RAFT and LoRA, enabling effective question answering in environments with limited computational resources.

Contribution

It proposes a novel combination of RAFT and LoRA to improve efficiency and reduce resource requirements for knowledge-intensive QA tasks.

Findings

CRAFT achieves comparable performance to larger models.

It significantly reduces fine tuning and storage needs.

Faster inference times are demonstrated in resource-constrained settings.

Abstract

Retrieval Augmented Generation (RAG) is a common method for integrating external knowledge into pretrained Large Language Models (LLMs) to enhance accuracy and relevancy in question answering (QA) tasks. However, prompt engineering and resource efficiency remain significant bottlenecks in developing optimal and robust RAG solutions for real-world QA applications. Recent studies have shown success in using fine tuning to address these problems; in particular, Retrieval Augmented Fine Tuning (RAFT) applied to smaller 7B models has demonstrated superior performance compared to RAG setups with much larger models such as GPT-3.5. The combination of RAFT with parameter-efficient fine tuning (PEFT) techniques, such as Low-Rank Adaptation (LoRA), promises an even more efficient solution, yet remains an unexplored area. In this work, we combine RAFT with LoRA to reduce fine tuning and storage requirements and gain faster inference times while maintaining comparable RAG performance. This results in a more compute-efficient RAFT, or CRAFT, which is particularly useful for knowledge-intensive QA tasks in resource-constrained environments where internet access may be restricted and hardware resources limited.