CodeBERT: A Pre-Trained Model for Programming and Natural Languages

TL;DR

CodeBERT is a bimodal pre-trained Transformer model that learns representations for programming and natural languages, enabling improved performance on code search and documentation tasks, and demonstrating knowledge transfer in zero-shot settings.

Contribution

It introduces CodeBERT, a novel bimodal pre-trained model for programming and natural languages, with a hybrid training objective that leverages both bimodal and unimodal data.

Findings

Achieves state-of-the-art results on code search and documentation generation.

Performs well in zero-shot NL-PL probing tasks.

Utilizes a hybrid training objective with replaced token detection.

Abstract

We present CodeBERT, a bimodal pre-trained model for programming language (PL) and nat-ural language (NL). CodeBERT learns general-purpose representations that support downstream NL-PL applications such as natural language codesearch, code documentation generation, etc. We develop CodeBERT with Transformer-based neural architecture, and train it with a hybrid objective function that incorporates the pre-training task of replaced token detection, which is to detect plausible alternatives sampled from generators. This enables us to utilize both bimodal data of NL-PL pairs and unimodal data, where the former provides input tokens for model training while the latter helps to learn better generators. We evaluate CodeBERT on two NL-PL applications by fine-tuning model parameters. Results show that CodeBERT achieves state-of-the-art performance on both natural language code search and code documentation generation tasks. Furthermore, to investigate what type of knowledge is learned in CodeBERT, we construct a dataset for NL-PL probing, and evaluate in a zero-shot setting where parameters of pre-trained models are fixed. Results show that CodeBERT performs better than previous pre-trained models on NL-PL probing.