English(EN) Model-Preserving Adaptive Rounding with YAQA

Together AI发布YAQA以改进LLM量化

作者 PulseAugur 编辑部 · [1 个来源] · 2025-06-05 00:00

Together AI推出了一种新颖的大型语言模型后训练量化技术YAQA。该方法旨在通过直接最小化KL散度来比现有算法更有效地保留原始模型的输出。YAQA通过近似KL散度的Hessian来实现这一点，与当前的舍入方法相比，KL散度降低了30%以上，并在下游任务上提高了性能。 AI

影响 YAQA改进的量化技术有望以最小的性能下降，更高效地部署大型语言模型。

排序理由该集群描述了来自AI研究组织的一篇新技术论文和方法发布。[lever_c_demoted from research: ic=1 ai=1.0]

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: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群描述了来自AI研究组织的一篇新技术论文和方法发布。[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, model release; 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: 492 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。