Fine-Tuning ChatGPT

I have been reading a few articles on how to fine tune a model following these resources https://towardsdatascience.com/unleashing-the-power-of-gpt-how-to-fine-tune-your-model-da35c90766c4, https://medium.com/@olahsymbo/fine-tuning-openai-gpt-3-to-build-custom-chatbot-fe2dea524561, and learning how to clean up data. Learning Devops practices for LLM was made simpler by listening to some of the talks from the LLM Bootcamp https://www.youtube.com/watch?v=Fquj2u7ay40

A majority of my time this week was spent on wrapping my head around all of these practices and continued learning in the AI space. I have begun to create a pipeline for processing the interviews from David Leech’s dataset related to a number of climate disasters that have been occuring in BC over the past few years. Right now my pipeline converts the text into a CSV file and then the CSV file into a JSONL file. I hope to actually get a chance to fine-tune a model using my data by the end of the week.

-Nick