Skip to content

Latest commit

 

History

History
34 lines (25 loc) · 705 Bytes

File metadata and controls

34 lines (25 loc) · 705 Bytes

Deep-Robust-Clustering-of-Source-Code

Implementation of the Deep Robust Clustering technique as presented in the paper Deep Robust Clustering by Contrastive Learning, but adapted for clustering source code.

Prerequisites

  • python3.5+
  • pytorch
pip install -r requirements.txt

Usage

Extract and augment methods from a directory of source code.

python extract_and_augment_methods.py -d <path to directory>

Run preprocessing to build vocabularies and datasets

./preprocess.sh

Edit config.py and then train the model

python train.py

Extract features from test set and visualize clusters

python feature_extraction.py