A ML-LLM framework for constructing a data augmentation pipeline by LLM-prompting.
This is the official code repository for the IRSE@FIRE 2023 track paper: "A ML-LLM pairing for better code comment classification".
This directory contains two data files:
- Seed data: The data provided by the task organizers to train the ML models.
- LLM-generated data: The data generated by a LLM assistant (ChatGPT in this case) to evaluate the overall increase in model performance after data augmentation.
This directory contains the code for training and evaluating ML models on both the seed and LLM data. The code also contains data augmentation techniques using synthetic data.
This directory contains the official results submitted to the shared task.