Evaluating How Fine-tuning on Bimodal Data Effects Code Generation
This addresses the need for more reliable code generation models for developers, though it is incremental as it builds on existing fine-tuning methods with new data.
The study tackled the problem of how fine-tuning language models on bimodal coding forum data affects code generation performance and reliability, finding that it improved average pass@k scores by 54.64% on HumanEval and 85.35% on Mostly Basic Program Problems while reducing syntax and runtime errors, but also revealed decreased runnability at higher temperatures.
Despite the increase in popularity of language models for code generation, it is still unknown how training on bimodal coding forums affects a model's code generation performance and reliability. We, therefore, collect a dataset of over 2.2M StackOverflow questions with answers for finetuning. These fine-tuned models have average $pass@k$ improvements of 54.64% and 85.35% on the HumanEval (Chen et al., 2021) and Mostly Basic Program Problems (Austin et al., 2021) tasks, respectively. This regime further decreases the number of generated programs with both syntax and runtime errors. However, we find that at higher temperatures, there are significant decreases to the model's ability to generate runnable programs despite higher $pass@k$ scores, underscoring the need for better methods of incorporating such data that mitigate these side effects. The code can be found https://github.com/gabeorlanski/bimodalcode-generation