One Adapter for All Programming Languages? Adapter Tuning for Code Search and Summarization

원문 초록

As pre-trained models automate many code intelligence tasks, a widely used paradigm is to fine-tune a model on the task dataset for each programming language. A recent study reported that multilingual fine-tuning benefits a range of tasks and models. However, we find that multilingual fine-tuning leads to performance degradation on recent models UniXcoder and CodeT5. To alleviate the potentially catastrophic forgetting issue in multilingual models, we fix all pre-trained model parameters, insert the parameter-efficient structure adapter, and fine-tune it. Updating only 0.6\% of the overall parameters compared to full-model fine-tuning for each programming language, adapter tuning yields consistent improvements on code search and summarization tasks, achieving state-of-the-art results. In addition, we experimentally show its effectiveness in cross-lingual and low-resource scenarios. Multilingual fine-tuning with 200 samples per programming language approaches the results fine-tuned with the entire dataset on code summarization. Our experiments on three probing tasks show that adapter tuning significantly outperforms full-model fine-tuning and effectively overcomes catastrophic forgetting.