https://github.com/codito/slm-finetune

https://github.com/codito/slm-finetune

Last synced: 3 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/codito/slm-finetune
• Owner: codito
• License: mit
• Created: 2024-08-21T18:38:28.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2024-09-23T16:18:00.000Z (over 1 year ago)
• Last Synced: 2025-01-07T06:41:50.860Z (about 1 year ago)
• Language: Jupyter Notebook
• Size: 2.41 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Small Language Models Finetuning

Explorations around finetuning the [Gemma 2](https://huggingface.co/collections/google/gemma-2-release-667d6600fd5220e7b967f315) models for learning and fun!

- [Function Calling](https://github.com/codito/gemma-2-ft/blob/master/gemma-2-2b-function-calling.ipynb) tries to create a finetuned model which can choose a tool given a natural language query. 

## Running the notebooks

We're using Kaggle's free GPU tier (30h/week) to develop these notebooks.

You can fork, or download and import the notebooks to Kaggle (from File -> Import Notebook).

Later [link](https://www.kaggle.com/discussions/product-feedback/295170) your Github account in Kaggle and save your changes. 

## License

MIT