arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2104.09694cs.CLcs.IRcs.LG
收藏

Transformers的有效预培训目标

Efficient pre-training objectives for Transformers

University of Trento(特伦托大学);Amazon Alexa(亚马逊Alexa)

Luca Di Liello, Matteo Gabburo, Alessandro Moschitti

摘要:

Transformer架构深刻地改变了自然语言处理,性能超过了以前所有最先进的模型。然而,众所周知的Transformer模型,如BERT、RoBERTa和GPT-2需要巨大的计算预算来创建高质量的上下文表示。在本文中,我们研究了几种有效的基于transformers的模型的预训练目标。通过在不同的任务上测试这些目标,我们确定了ELECTRA模型的哪些新功能是最相关的。我们确认,当输入不包含掩码令牌时,Transformers预训练得到改进,并且使用整个输出来计算损失减少了训练时间。此外,受ELECTRA的启发,我们研究了一个由两个模块组成的模型;一个基于统计模型的发电机和一个简单的发电机,对计算性能没有影响。此外,我们证明了在损失计算过程中消除MASK令牌并考虑整个输出是提高性能的必要选择。此外,我们表明,它是可以有效地训练BERT类模型使用的歧视性的方法,在ELECTRA,但没有一个复杂的发电机,这是昂贵的。最后,我们表明,ELECTRA受益匪浅,从一个国家的最先进的超参数搜索。

英文摘要:

The Transformer architecture deeply changed the natural language processing, outperforming all previous state-of-the-art models. However, well-known Transformer models like BERT, RoBERTa, and GPT-2 require a huge compute budget to create a high quality contextualised representation. In this paper, we study several efficient pre-training objectives for Transformers-based models. By testing these objectives on different tasks, we determine which of the ELECTRA model's new features is the most relevant. We confirm that Transformers pre-training is improved when the input does not contain masked tokens and that the usage of the whole output to compute the loss reduces training time. Moreover, inspired by ELECTRA, we study a model composed of two blocks; a discriminator and a simple generator based on a statistical model with no impact on the computational performances. Besides, we prove that eliminating the MASK token and considering the whole output during the loss computation are essential choices to improve performance. Furthermore, we show that it is possible to efficiently train BERT-like models using a discriminative approach as in ELECTRA but without a complex generator, which is expensive. Finally, we show that ELECTRA benefits heavily from a state-of-the-art hyper-parameters search.

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。

↑