https://github.com/evilfreelancer/chatgpt-notebooks
Collection of small notebooks for pretraining ChatGPT
https://github.com/evilfreelancer/chatgpt-notebooks
Last synced: 2 months ago
JSON representation
Collection of small notebooks for pretraining ChatGPT
- Host: GitHub
- URL: https://github.com/evilfreelancer/chatgpt-notebooks
- Owner: EvilFreelancer
- Created: 2023-08-23T19:05:17.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-08-26T12:38:21.000Z (over 2 years ago)
- Last Synced: 2025-04-04T03:41:24.390Z (9 months ago)
- Language: Jupyter Notebook
- Size: 7.81 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# ChatGPT-3.5 fine-tune notebook
Welcome, programmers!
This is your one-stop solution for fine-tuning the ChatGPT-3.5 Turbo model.
Contained within are essential resources to kickstart your model training journey: a sample dataset to give you a
glimpse into the training data structure, a handy cURL script that lets you initiate training without diving into Python
code, and a Jupyter notebook detailing the fine-tuning process
Whether you're an AI enthusiast or a seasoned practitioner, we've got you covered. Dive in and empower your ChatGPT!
## Related Blog Post
For a comprehensive walkthrough of the fine-tuning process and insights into the methodologies used, refer to
this [blog post](https://dzen.ru/a/ZOZQNS_FBUJZ2tjU) (Note: The blog post is in Russian).
## Requirements
* Python 3.10 or later
* Python VirtualEnv
* Jupyter Notebook
* [RecipeNLG](https://recipenlg.cs.put.poznan.pl/dataset) dataset placed in the root of this project.
* Required Python libraries (as listed in requirements.txt)
## Instructions
1. Ensure you have all the requirements installed and the RecipeNLG dataset available.
2. There are specific cells in the notebook where you can input values or parameters for fine-tuning. Ensure to provide
appropriate values when prompted.
3. Run each cell in notebook sequentially to initiate the fine-tuning process.
4. The notebook will provide the fine-tuned model and display performance metrics after the process completes.
## Links
* https://dzen.ru/a/ZOZQNS_FBUJZ2tjU
* https://recipenlg.cs.put.poznan.pl/dataset
* https://openai.com/blog/gpt-3-5-turbo-fine-tuning-and-api-updates
* https://openai.com/pricing