This repo has multiple directories as below:
It fine tunes the hugging face models (eg llama2, BERT, Falcon, etc.) and is made up of a easy setup as well as customized setup file.
This directory fine tunes llama2-7b-chat (LLAMA2), nous-hermes2 (LLAMA2), bloom-560m (Bloom) by virtue of gradient ai library that has simplified some steps in the fine tunning.
Openai GPT directory is consisted of source code of fine tuning open ai models (GPT4, GPT3.5, etc.) as well as a simple dataset in jsonl format.