English(EN) PrePARE: Pre-AA Token Pruning for Frozen Multi-View Geometry Transformers

新的PrePARE方法大幅降低多视图几何Transformer的内存使用量

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-01 04:00

研究人员开发了PrePARE，一种通过在Token进入交替注意力（AA）堆栈之前对其进行剪枝来优化多视图几何Transformer的新颖方法。这种方法显著减少了内存使用量并提高了处理速度，使得像VGGT和MapAnything这样的复杂模型可以在性能较低的硬件上运行。PrePARE通过仅训练一个Token Scorer和一个Feature-guided Restoration模块来实现这一点，而核心AA堆栈保持冻结状态，在ScanNetv2等数据集上展示了显著的内存节省和性能提升。 AI

影响该方法有望在标准硬件上实现更高效的复杂多视图几何模型的训练和部署。

排序理由这是一篇详细介绍优化现有Transformer模型新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 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: 18 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 这是一篇详细介绍优化现有Transformer模型新方法的学术论文。[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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CV TIER_1 English(EN) · Haotang Li, Zhenyu Qi, Shaohan Henry Wang, Kebin Peng, Zi Wang, Qing Guo, Sen He, Huanrui Yang · 2026-10-01 04:00

PrePARE：冻结多视图几何 Transformer 的预 AA Token 修剪

arXiv:2605.08371v2 Announce Type: replace Abstract: Multi-view geometry transformers are feed-forward 3D foundation models that jointly predict depth maps, point maps, and camera poses for N images in a single forward pass. Most of them build on an alternating-attention (AA) stac…

本文如何被排名

报道来源 [1]

PrePARE：冻结多视图几何 Transformer 的预 AA Token 修剪

相关实体

相关话题