This document summarizes my learnings and experiences with fine-tuning Large Language Model models. The goal is to create a concise, revisitable resource that simplifies understanding and implementation of fine-tuning techniques for LLMs.
Currently, I only summarize and note down Fine-tuning techniques for LLMs without detailed implementation.
You can check this awesome paper/book The Ultimate Guide to Fine-Tuning LLMs from Basics to Breakthroughs to get more information about fine-tuning technique. Also check this one Instruction Tuning Survey to understand more about Instruction Tuning.
- Insight From 100 Experiments - Finetuning LLMs with LoRA and QLoRA: Insights from Hundreds of Experiments
- Advanced Guide to training a LoRA - Essential to Advanced Guide to training a LoRA
- Efficient Training on a Single GPU - Methods and tools for efficient training on a single GPU