Yes! Sort of. There are free code libraries and cloud computing resources which enable pretty much anyone with some data and decent internet access to customize language models…if they have the necessary programming and machine learning experience.

I’m hoping to narrow that gap with this project. While you’re still going to need some coding experience, I’m providing code notebooks that will make it easy for anyone who knows a little Python to mimic my methods.

Fine-tuning: https://www.kaggle.com/code/sarahimmel/data-doppelganger-fine-tuning-tutorial

RAG: https://www.kaggle.com/code/sarahimmel/datadoppelganger-rag-prompting-tutorial

You can also reach out via email (s.immel@ed.ac.uk) or LinkedIn with any questions or comments.