DeCodR

Semantic Code & Repository Search - Project Report

Datasets: Repository Search, Code Search

This was a final group project for CS 685 (Advanced NLP at UMASS Amherst)

Project Structure

• code_lm_pretraining contains the code we used for MLM pre-training
• code_search has the model implementations for the code search task on the Stack Overflow dataset
• repo_search has the model implementations for the repository search task on the Github dataset