What is the best way to create a fine-tuning model to deal with SQL?

Hello! I need your help for advice.

I need to create a fine-tuning model so that it has knowledge of my SQL tables and their fields.

He will be responsible and help users find information and generate SQL queries.

My question is: How and what is the easiest and most accurate way to create the training model? Remembering that there are several tables and several fields.

Thanks in advance!


Fine tuning is not an ideal way to give the model new information, it trains the model on new patterns rather than explicit data. You may want to look into Embeddings as a method in increase the factual information the model can process, you can find information here OpenAI Platform