English(EN) LSTM-Style Looped Transformers with Shared KV Caches

LSTM风格的Transformer减少了参数和缓存需求

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-28 04:00

研究人员开发了一种LSTM风格的Transformer，它在循环核心中重用了权重并共享了键值缓存，显著减少了参数数量和缓存需求。这种方法可以用更少的参数和更少的内存实现深度计算，在数据和模型存储有限的情况下尤其有利。训练方法通过一致性损失解决了并行和递归缓存条目之间的不匹配问题，提高了在不相关文档和各种模型宽度上的性能。 AI

影响这项研究可能带来更高效的Transformer模型，降低深度学习任务的计算和内存需求。

排序理由详细介绍新型模型架构的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 13 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍新型模型架构的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, infra; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.LG TIER_1 English(EN) · Andreas Kirsch · 2026-10-28 04:00

LSTM风格的循环Transformer与共享KV缓存

arXiv:2610.33360v1 Announce Type: new Abstract: Recurrent transformers reuse weights to perform deep computation with fewer parameters. We study an LSTM-style transformer that also shares one key--value cache across its recurrent core. At width 1024, it uses 40.4% fewer parameter…

本文如何被排名

报道来源 [1]

LSTM风格的循环Transformer与共享KV缓存

相关实体

相关话题