FedRAG: A Framework for Fine-Tuning Retrieval-Augmented Generation Systems

TL;DR

FedRAG is a versatile framework that enables effective fine-tuning of retrieval-augmented generation systems in both centralized and federated settings, enhancing their adaptability and performance.

Contribution

It introduces a comprehensive framework supporting state-of-the-art fine-tuning methods for RAG systems, with seamless transition between centralized and federated training.

Findings

Supports state-of-the-art fine-tuning methods

Enables seamless conversion from centralized to federated training

Deep integration with the modern RAG ecosystem

Abstract

Retrieval-augmented generation (RAG) systems have been shown to be effective in addressing many of the drawbacks of relying solely on the parametric memory of large language models. Recent work has demonstrated that RAG systems can be improved via fine-tuning of their retriever and generator models. In this work, we introduce FedRAG, a framework for fine-tuning RAG systems across centralized and federated architectures. FedRAG supports state-of-the-art fine-tuning methods, offering a simple and intuitive interface and a seamless conversion from centralized to federated training tasks. FedRAG is also deeply integrated with the modern RAG ecosystem, filling a critical gap in available tools.