Code to Fine-tune ChatGPT w/ synthetic GPT-4 dataset

A Jupyter Notebook in Python to fine-tune ChatGPT. Plus a synthetic dataset will be created by GPT-4 for a user specific task. Multi GPT system interaction and advanced prompt engineering. Plus Matt Shumer's GPT-LLM-Trainer explained in detail.
All rights with Matt Shumer for his Jupyter NB on fine-tuning LLama 2 model:
colab.research.google.com/dri...
See also Matt Shumer's Github repo for the GPT-LLM-Trainer:
github.com/mshumer/gpt-llm-tr...
Recommend: Also check out this informative YT video by ‪@SophiaYangDS‬ for her detailed code implementation on how to fine-tune ChatGPT with a different dataset from HuggingFace:
• Fine-tune GPT3.5 Turbo...
#ai
#chatgptprompts
#code

[@PoornimaDevi-yx9oh]

Cool work & showcasing ! Thanks for sharing this here.

[@echofloripa]

I did a test using questions and answers about the elections process in Brazil. It had 67 questions and answers. I tried the default 3 epochs, 5, 7 and even 12. In none of the cases I managed to get the same response I had trained on, for exact same system message and user message. I tried in Portuguese and English language, and the result was the same. Yes, it gave a different response compared to the base model, but yet, never a correct answer. For the English dataset test I trimmed the 67 questions to only 10. You can check the loss of the training using its api and the numbers was erratic. I guess that at least in gpt3.5-turbo fine tuning, it's not possible to get it increase it's knowledge. I did some tests with open-source llms, but I still have to train with llama2. Maybe fine-tuning isn't really fit for that, and you have to use embeddings and vector databases to achieve that.

[@sameekhan1936]

excited to try this out!

[@yannickpezeu3419]

Thanks a lot for all your work, I have a question: All the fine tuning I see is done with a dataset of question answers. Is it possible to give the model some raw text ? I wish to give the model all the scientific publications and courses of my institution (EPFL) So that the model would know everything about the research and courses done at EPFL. In an ideal world I would then be able to ask it stuff about who is expert in which field, what is the best course to learn this or that. Have you heard of something like this ?

[@yannickpezeu3419]

thx

[@cecilsalas8721]

🤩👍

[@code4AI]

Great to hear!