arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.11972cs.LGcs.CL
收藏

Transformer修改是否在实施和应用程序之间传输?

Do Transformer Modifications Transfer Across Implementations and Applications?

Google(谷歌)
Sharan Narang, Hyung Won Chung, Yi Tay, William Fedus, Thibault Fevry, Michael Matena, Karishma Malkan, Noah Fiedel, Noam Shazeer, Zhenzhong Lan, Yanqi Zhou, We… 展开作者

Sharan Narang, Hyung Won Chung, Yi Tay, William Fedus, Thibault Fevry, Michael Matena, Karishma Malkan, Noah Fiedel, Noam Shazeer, Zhenzhong Lan, Yanqi Zhou, Wei Li, Nan Ding, Jake Marcus, Adam Roberts, Colin Raffel

摘要:

自从Transformer体系结构在三年多前引入以来,研究界已经提出了大量的修改,其中相对较少的被广泛采用。在本文中,我们全面评估了许多这些修改在一个共享的实验设置,涵盖了Transformer在自然语言处理中的大多数常见用途。令人惊讶的是,我们发现大多数修改并没有有意义地提高性能。此外,我们发现的大多数有益的Transformer变体要么是在我们使用的相同代码库中开发的,要么是相对较小的变化。我们推测,性能的改善可能强烈依赖于实现细节,并相应地提出了一些建议,以提高实验结果的一般性。

英文摘要:

The research community has proposed copious modifications to the Transformer architecture since it was introduced over three years ago, relatively few of which have seen widespread adoption. In this paper, we comprehensively evaluate many of these modifications in a shared experimental setting that covers most of the common uses of the Transformer in natural language processing. Surprisingly, we find that most modifications do not meaningfully improve performance. Furthermore, most of the Transformer variants we found beneficial were either developed in the same codebase that we used or are relatively minor changes. We conjecture that performance improvements may strongly depend on implementation details and correspondingly make some recommendations for improving the generality of experimental results.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑