Computer Science > Software Engineering
[Submitted on 29 Jun 2022 (this version), latest version 21 Nov 2022 (v5)]
Title:Diet Code is Healthy: Simplifying Programs for Pre-Trained Models of Code
View PDFAbstract:Pre-trained code representation models such as CodeBERT have demonstrated superior performance in a variety of software engineering tasks, yet they are often heavy in complexity, quadratically with the length of the input sequence. Our empirical analysis of CodeBERT's attention reveals that CodeBERT pays more attention to certain types of tokens and statements such as keywords and data-relevant statements. Based on these findings, we propose DietCodeBERT, which aims at lightweight leverage of large pre-trained models for source code. DietCodeBERT simplifies the input program of CodeBERT with three strategies, namely, word dropout, frequency filtering, and an attention-based strategy which selects statements and tokens that receive the most attention weights during pre-training. Hence, it gives a substantial reduction in the computational cost without hampering the model performance. Experimental results on two downstream tasks show that DietCodeBERT provides comparable results to CodeBERT with 40% less computational cost in fine-tuning and testing.
Submission history
From: Xiaodong Gu [view email][v1] Wed, 29 Jun 2022 04:04:38 UTC (742 KB)
[v2] Wed, 17 Aug 2022 09:23:19 UTC (1,463 KB)
[v3] Tue, 30 Aug 2022 12:14:22 UTC (1,463 KB)
[v4] Tue, 20 Sep 2022 08:18:43 UTC (1,498 KB)
[v5] Mon, 21 Nov 2022 13:31:39 UTC (1,498 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.