Fine tuning defog/llama-3-sqlcoder-8b

#16
by kchakkarwar - opened

Hi defog Team,
I am using defog/llama-3-sqlcoder-8b LLM model for our one of the usecase. We are using Microsoft sql server management tool db to fetch the query generated and we found that for most of the nlp query , model is not generating sql query properly.
So we decided to fine tune the model as per our db schema in order to generate correct sql query by model.

Can you please help share how to fine tune the model? Also can you share sample dataset format for NLP query that you used for training llama-3-sqlcoder-8b? This sample dataset will help us to get an idea what questions need to use for training?

Hi @jp-defog ,
Any update on this ? Please help

Defog.ai org

@kchakkarwar I'm sorry, we don't provide explicit instructions or data on how to finetune the model.

jp-defog changed discussion status to closed

Sign up or log in to comment