Code-Optimise: Self-Generated Preference Data for Correctness and Efficiency
This addresses the challenge of balancing correctness and efficiency in code generation for developers, though it is incremental as it builds on existing optimization methods.
The paper tackles the problem of code language models generating correct but inefficient solutions by introducing Code-Optimise, a framework that uses self-generated preference data to optimize for both correctness and runtime, achieving a 6% runtime reduction on in-domain data and up to 48% shorter solutions on benchmarks.
Code Language Models have been trained to generate accurate solutions, typically with no regard for runtime. On the other hand, previous works that explored execution optimisation have observed corresponding drops in functional correctness. To that end, we introduce Code-Optimise, a framework that incorporates both correctness (passed, failed) and runtime (quick, slow) as learning signals via self-generated preference data. Our framework is both lightweight and robust as it dynamically selects solutions to reduce overfitting while avoiding a reliance on larger models for learning signals. Code-Optimise achieves significant improvements in pass@k while decreasing the competitive baseline runtimes by an additional 6% for in-domain data and up to 3% for out-of-domain data. As a by-product, the average length of the generated solutions is reduced by up to 48% on MBPP and 23% on HumanEval, resulting in faster and cheaper inference. The generated data and codebase is open-sourced at https://github.com/huawei-noah/HEBO/tree/Code_Optimise.