In this article I tune a base LLama2 LLM to output SQL code. I use Parameter Efficient Fine-Tuning Techniques to optimise the process.
Originally appeared here:
How to efficiently fine-tune your own open-source LLM using novel techniques — code provided