https://github.com/k-dizzled/ml4se-program-repair

https://github.com/k-dizzled/ml4se-program-repair

Last synced: about 2 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/k-dizzled/ml4se-program-repair
• Owner: K-dizzled
• License: lgpl-2.1
• Created: 2024-11-09T13:39:41.000Z (about 2 months ago)
• Default Branch: main
• Last Pushed: 2024-11-09T13:53:51.000Z (about 2 months ago)
• Last Synced: 2024-11-09T14:35:10.389Z (about 2 months ago)
• Language: TypeScript
• Size: 0 Bytes
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# Program Repair Project

Utilizing Large Language Models to preair a dataset and use it to train a model to repair code. This particular `ts` package generates solutions for coding problems from the `livecodebench` dataset via some small model, then, stress tests solutions using data provided in the dataset and in case of incorrect behavior -- tries to fix solutions using the execution feedback and a bigger model. Via this procedure a dataset is created. 

Authors: *Nikita Khramov*, *Andrei Kozyrev* and *Ivan Kabashnyi*

ML4SE, Constructor University, Bremen, Autumn 2024

## Prerequisites

Setup the npm version via `nvm`:

```bash

nvm use

```

Install dependencies:

```bash

npm install

```

## Development

Run the entry point:

```bash

npm run serve

```

Run tests:

```bash

npm test

```