https://github.com/abdulhadi806/llm_fune_tuning_hackathon
In the recent competition, we were challenged to finetune a model that can convert a LaTeX expressions into Python code effectively. My team, which I led, secured 6th place overall.
https://github.com/abdulhadi806/llm_fune_tuning_hackathon
ai fine-tuning h100 hackathon llm
Last synced: about 1 year ago
JSON representation
In the recent competition, we were challenged to finetune a model that can convert a LaTeX expressions into Python code effectively. My team, which I led, secured 6th place overall.
- Host: GitHub
- URL: https://github.com/abdulhadi806/llm_fune_tuning_hackathon
- Owner: AbdulHadi806
- Created: 2024-08-30T16:37:54.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-08-30T17:06:12.000Z (almost 2 years ago)
- Last Synced: 2025-02-06T23:31:10.807Z (over 1 year ago)
- Topics: ai, fine-tuning, h100, hackathon, llm
- Language: Jupyter Notebook
- Homepage:
- Size: 12.3 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Convert Latex to Python code
We fine-tuned a Salesforce/CodeT5p-770 on a 35k dataset and gain an accuracy of 74% on test data scoring 6th position. We applied Lora and post quanitization to speeding up
the inference and training. Moreover we also applied post processing in order to fix some common indentation issues
## Space to test out our model inference
[Huggingface space](https://huggingface.co/spaces/AbdulHadi806/LaTeX-to-Python-CodeT5-Finetune)
[Model](https://huggingface.co/AbdulHadi806/codeT5-finetuned-LaTexToPythonCode-30kDataset)
## My Team
[AbdulHadi806](https://github.com/AbdulHadi806)
[acroash1122](https://github.com/acroash1122)
[abdulsaboor728](https://github.com/abdulsaboor728)
[ashawesomeone](https://github.com/ashawesomeone/)