Introduction

SolBert

Usage

1. Install the required packages Run the following commands: pip install -r requirements.txt

2. Download the corresponding dataset： Link to experimental dataset:：https://drive.google.com/drive/folders/1vauZV2VbTkPSzDG9MM6E1O06Iaad_3Rl?usp=drive_link Experimentally constructed datasets: clone, bug and cluster datasets are placed under code/experiment/clone(bug|cluster)/data, respectively.

3. Training model and parameters: The pre-training model can be downloaded at: https://drive.google.com/drive/folders/1ZXg2r2xideI9w3bSeMicH4X2Gon714Bo?usp=drive_link pre-training model：solBert(code/solidity_model) mirror_soliBert: (code/mirror_bert/tmp/mirror_bert_mean/) bert_whitening:(code/bertWhitening/data/) mirror_bert_whitening:(code/mirror_bert_whitening/data/)

4. Experiment 1.modify code/config.py path
   2.modify code/experiment/config.py path

   Clone

   step1 Build clone experimental dataset
   step2 to step5 use different methods respectively:
       python step2_bert_clone_detection.py avg_first_last (Clone experiment with avg_first_last pooling, the rest of the steps are similar)
   

   Cluster

   Similar to the clone experiment
   

   Bug

   Similar to the clone experiment