CodeGeeX: A Pre-Trained Model for Code Generation with Multilingual Benchmarking on HumanEval-X

TL;DR

CodeGeeX is a large, multilingual code generation model trained on 850 billion tokens, outperforming similar models on code tasks and integrated into developer tools to enhance coding efficiency.

Contribution

Introduces CodeGeeX, a 13-billion-parameter multilingual code generation model trained on extensive data, with new benchmark HumanEval-X and practical extensions for developer tools.

Findings

Outperforms similar scale models on code generation and translation tasks

Develops the HumanEval-X benchmark for multilingual evaluation

Enhances developer productivity through CodeGeeX integrations

Abstract

Large pre-trained code generation models, such as OpenAI Codex, can generate syntax- and function-correct code, making the coding of programmers more productive and our pursuit of artificial general intelligence closer. In this paper, we introduce CodeGeeX, a multilingual model with 13 billion parameters for code generation. CodeGeeX is pre-trained on 850 billion tokens of 23 programming languages as of June 2022. Our extensive experiments suggest that CodeGeeX outperforms multilingual code models of similar scale for both the tasks of code generation and translation on HumanEval-X. Building upon HumanEval (Python only), we develop the HumanEval-X benchmark for evaluating multilingual models by hand-writing the solutions in C++, Java, JavaScript, and Go. In addition, we build CodeGeeX-based extensions on Visual Studio Code, JetBrains, and Cloud Studio, generating 4.7 billion tokens for tens of thousands of active users per week. Our user study demonstrates that CodeGeeX can help to increase coding efficiency for 83.4% of its users. Finally, CodeGeeX is publicly accessible and in Sep. 2022, we open-sourced its code, model weights (the version of 850B tokens), API, extensions, and HumanEval-X at https://github.com/THUDM/CodeGeeX.