Code for the paper: "[Parameter Efficient Fine-Tuning of Pre-trained Code Models for Just-in-Time Defect Prediction]()""

This is the official repository for our paper on utilizing pre-trained code lanugage models for the task of Just-in-Time Software Defect Prediction.

The repository includes:

• The original datasets.
• Code to pre-process and generate our versions of the datasets.
• Code for visualizations in the paper and code for generating custom tokenizers
• Code to train (fine-tune) and evaluate the models presented in the paper using the proposed fine-tuning techniques.
• Code to generate custom datasets for CC2Vec.
• CC2Vec training and evaluation scripts.