Skip to content

julialfk/CompBERT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

58 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CompBERT

This repository contains the implementation of a language model, named CompBERT, designed to facilitate the assessment of functional completeness in software projects. CompBERT helps users identify which methods or functions within a codebase are likely related to a given natural language (NL) feature description. This project builds upon Microsoft's UniXcoder model and code search approach for fine-tuning.

Key Features

  • Custom dataset creation: The project includes a method for creating datasets from IlmSeven and SEOSS open-source projects.
  • Fine-tuning datasets: The full training, development and evaluation datasets constructed and used for this project can be found in a separate HuggingFace repository.
  • Fine-tuning: The scripts and configurations used for fine-tuning the UniXcoder model to create CompBERT.
  • Output models: The models created used in the experiments of this thesis can be found in a separate HuggingFace repository. These models are fine-tuned to score each method, indicating its relevance to a given NL feature description.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors