Finetuning Llama 2.7B model using open weights of Llama from Hugging face and the dataset called guanico Thanks to Krishnaik for his tutorials on youtube.
This is a project on finetuning and not focused on RAG. Download/ clone the git repository and upload the jupyter notebook to google collab. Use the free version GPU for faster operations. If you are running on your local machine it should have atleast 32GB of RAM and good enough GPU as Llama 2.7B is a very huge model to run locally. Fine tuning takes couple minutes depending on the GPU provisioned for operation. I have downloaded the finetuned model into my drive and used streamlit as front end to launch as a chatbot in other repository.