English(EN) TileShare: Training-Free Acceleration of Overlapping Tiled ViT Decoding

TileShare 无需重新训练即可加速 ViT 解码

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-26 04:00

研究人员开发了 TileShare，一种无需额外训练即可加速 Vision Transformer (ViT) 模型解码过程的新颖方法。该技术解决了固定窗口 ViT 解码器中的计算冗余问题，其中重叠的切片会导致重复处理。通过将重叠 token 的结果智能地融合到共享状态中，并为每个切片选择性地保留几层，TileShare 显著减少了计算量。该方法在各种视频和图像数据集上实现了更快的解码速度，使用 FP8 时最高可达 1.96 倍，同时保持了可比或更优的性能指标，如 PSNR、LPIPS 和 DISTS。 AI

影响提供了一种无需训练即可提高 Vision Transformer 模型推理速度的方法，有望降低 AI 应用的计算成本。

排序理由该条目是 arXiv 预印本，详细介绍了一种加速现有模型的新方法。[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; 该条目是 arXiv 预印本，详细介绍了一种加速现有模型的新方法。[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) · Xiangzhong Ye, Zhongshen Li, Shen You, Ziyang Meng, Liang Xu, Junkai Ji, Qiuzhen Lin, Xiangtao Li, Ka-Chun Wong · 2026-10-26 04:00

TileShare：无需训练即可加速重叠瓦片 ViT 解码

arXiv:2610.32233v1 Announce Type: new Abstract: Fixed-window Vision Transformer (ViT) decoders run at deployment resolution as overlapping tiles: every token in an overlap band is processed by several full decoder instances. In the 2.4B-parameter MiniMax-H3 video decoder, about h…

本文如何被排名

报道来源 [1]

TileShare：无需训练即可加速重叠瓦片 ViT 解码

相关实体

相关话题