Skip to content

RAG-FiT

Fine tune LLMs on RAG enhanced data

Visit official site

What RAG-FiT is

The library is described as a tool for improving LLM use of external information through specially created RAG enhanced datasets.

What you can do with it

  • Build a RAG enhanced dataset
  • Train a model with PEFT
  • Run inference and evaluation

Application scenarios

  • A PubMedQA run aids medical questions
  • A news summary task gains relevance
  • Legal text is used for fine tuning

Pros & cons

✓ What we like

  • Config files customise the workflow
  • Hugging Face Transformers and OpenAI are supported

! What to watch out for

  • A RAG dataset is created before training
  • The library centres on RAG workflows

FAQ

What training method is named?

The feature list names parameter efficient fine tuning.

Which metrics are listed?

The feature list names EM, F1 and ROUGE.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review