Our senior comps project for Carleton College.
This repository contains code for implementing, testing, analyzing, and applying different methods of code summarization and retrieval.
- BVAE: Bimodal Variational AutoEncoder. An experimental model architecture proposed by Chen and Zhou (2018) that is used for both retrieval and summarization
- Baselines:
- Retrieval: RET-IR
- Summarization: IR
- Transformer (experimental, for summarization only)
Trained BVAE models in TensorFlow SavedModel format can be found in models/.