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 2406.11753cs.CLcs.LG
收藏

一种面向语言模型计算高效微调的语义感知层冻结方法

A Semantic-Aware Layer-Freezing Approach to Computation-Efficient Fine-Tuning of Language Models

Monash University(莫纳什大学);Technical University of Munich(慕尼黑工业大学);Chongqing University(重庆大学)

Jian Gu, Aldeida Aleti, Chunyang Chen, Hongyu Zhang

AI总结:

针对语言模型微调成本高且现有方法忽略“在哪里微调”的问题,提出一种基于语义感知的层冻结方法,通过分析潜在表示转移轨迹和缩放定律评估各层增益,有效缩小微调范围并提升效率,且与PEFT等技术正交。

摘要:

微调语言模型(LMs)对于将模型适配到下游数据和任务至关重要。然而,全量微调通常成本高昂。现有工作,如参数高效微调(PEFT),通常关注“如何微调”,却忽略了“在哪里微调”的问题。作为通过回答在哪里微调来降低反向传播成本(在层级别)的先驱性工作,我们对LM推理过程进行了语义分析。我们首先提出使用潜在表示的转移轨迹来计算偏差(或损失)。然后,利用推导出的缩放定律公式,我们估计了每一层在减少偏差(或损失)方面的增益。进一步,我们缩小了微调的范围,并研究了LM微调的成本效益平衡。我们在知名的LMs和数据集上进行了广泛实验。结果表明,我们的方法有效且高效,并优于现有基线。我们的方法与诸如PEFT方法等其他提升微调效率的技术正交,为LM微调提供了实用价值。

英文摘要:

Finetuning language models (LMs) is crucial for adapting the models to downstream data and tasks. However, full finetuning is usually costly. Existing work, such as parameter-efficient finetuning (PEFT), often focuses on \textit{how to finetune} but neglects the issue of \textit{where to finetune}. As a pioneering work on reducing the cost of backpropagation (at the layer level) by answering where to finetune, we conduct a semantic analysis of the LM inference process. We first propose using transition traces of the latent representation to compute deviations (or loss). Then, using a derived formula of scaling law, we estimate the gain of each layer in reducing deviation (or loss). Further, we narrow down the scope for finetuning, and also, study the cost-benefit balance of LM finetuning. We perform extensive experiments across well-known LMs and datasets. The results show that our approach is effective and efficient, and outperforms the existing baselines. Our approach is orthogonal to other techniques for improving finetuning efficiency, such as PEFT methods, offering practical values on LM finetuning.

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。

↑