Function-to-Style Guidance of LLMs for Code Translation

TL;DR

F2STrans is a two-stage guiding paradigm for LLM-based code translation that enhances correctness and readability by combining functional and style learning, supported by a new comprehensive benchmark.

Contribution

The paper introduces F2STrans, a novel function-to-style guiding approach for LLMs in code translation, along with a new benchmark for evaluation.

Findings

F2STrans significantly improves code translation quality.

Qwen-1.5B outperforms larger models like GPT-4 on the benchmark.

The approach effectively balances correctness and readability in translated code.

Abstract

Large language models (LLMs) have made significant strides in code translation tasks. However, ensuring both the correctness and readability of translated code remains a challenge, limiting their effective adoption in real-world software development. In this work, we propose F2STrans, a function-to-style guiding paradigm designed to progressively improve the performance of LLMs in code translation. Our approach comprises two key stages: (1) Functional learning, which optimizes translation correctness using high-quality source-target code pairs mined from online programming platforms, and (2) Style learning, which improves translation readability by incorporating both positive and negative style examples. Additionally, we introduce a novel code translation benchmark that includes up-to-date source code, extensive test cases, and manually annotated ground-truth translations, enabling comprehensive functional and stylistic evaluations. Experiments on both our new benchmark and existing datasets demonstrate that our approach significantly improves code translation performance. Notably, our approach enables Qwen-1.5B to outperform prompt-enhanced Qwen-32B and GPT-4 on average across 20 diverse code translation scenarios.