🤖 AI Summary
This study investigates the applicability and efficacy of parameter-efficient fine-tuning (PEFT) methods for low-resource programming languages—specifically R—on large code language models. We empirically evaluate LoRA and Compacter on R code understanding and generation tasks using CodeT5 and CodeLlama as backbones. To our knowledge, this is the first work to demonstrate that PEFT achieves performance nearly on par with full fine-tuning (average gap <2.1%) while reducing GPU memory consumption by 68% and training time by 53%. Our findings confirm the effectiveness of knowledge transfer from natural-language pre-trained models to R-specific coding tasks and validate cross-lingual adaptation feasibility. The results extend the support boundary of large language models to niche programming languages and provide a robust pathway for lightweight deployment of code models in resource-constrained environments.
📝 Abstract
Recently, Large Langauge Models (LLMs) have gained a lot of attention in the Software Engineering (SE) community. LLMs or their variants pre-trained on code are used for many SE tasks. A main approach for adapting LLMs to the downstream task is to fine-tune the models. However, with having billions-parameters-LLMs, fine-tuning the models is not practical. An alternative approach is using Parameter Efficient Fine Tuning (PEFT), in which the model parameters are frozen and only a few added parameters are trained. Though the LLMs are used for programming languages such as Python and Java widely, their capability for low-resource languages is limited. In this work, we empirically study PEFT methods, LoRA and Compacter, on CodeT5 and CodeLlama. We will assess their performance compared to fully fine-tuned models, whether they can be used for knowledge transfer from natural language models to code (using T5 and Llama models), and their ability to adapt the learned knowledge to an unseen language. For the unseen language, we aim to study R, as it has a wide community. The adaptability with less computational costs makes LLMs accessible in scenarios where heavy computational resources are not available. Moreover, studying R opens new opportunities for using LLMs for other languages. We anticipate our findings to showcase the capabilities of PEFT for code LLMs for R and reveal the improvement areas.