研究人员开发了PaDoc，一种新颖的、基于布局的解析器，旨在提高文档解析的效率。与传统的顺序处理内容的端到端解析器不同，PaDoc将文档布局视为共享表示上的分支结构。这使得区域内容能够并行解码，显著降低处理深度并提高速度。PaDoc在OmniDocBench Full基准测试中取得了高精度，在文本编辑和公式识别方面优于其他端到端解析器，同时在大文档上还展示了卓越的吞吐量和更低的延迟。 AI
影响 这种并行解码方法可以显著加快AI系统中文档的处理速度，从而实现更快的信息提取和分析。
排序理由 该集群描述了一篇详细介绍新颖文档解析方法的最新研究论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2608.06146v1 Announce Type: new Abstract: End-to-end document parsers provide a unified interface, but serialize page layouts and regional contents into one autoregressive sequence. This formulation forces independent regions onto a decoding path whose length grows with the…
End-to-end document parsers provide a unified interface, but serialize page layouts and regional contents into one autoregressive sequence. This formulation forces independent regions onto a decoding path whose length grows with the total content, whereas crop-based two-stage par…
End-to-end document parsers provide a unified interface, but serialize page layouts and regional contents into one autoregressive sequence. This formulation forces independent regions onto a decoding path whose length grows with the total content, whereas crop-based two-stage par…