提出并行解码文档解析框架，提升速度与精度。
PaDoc: Layout-Grounded Parallel Decoding for Document Parsing
- 以布局为分支结构，实现内容并行解码。
- 在OmniDocBench上达91.1的布局F1和94.24的综合得分。
- 适合追求高并发、低延迟的端到端文档解析场景。
端到端文档解析器提供统一接口，但将页面布局和区域内容序列化为单一自回归序列，导致解码路径随内容总量增长而变长；基于图像裁剪的两阶段解析器虽暴露区域级并行性，却带来重复视觉预填充和碎片化页面上下文。为保留完整页面上下文同时消除依赖，我们提出PaDoc：一种以预测布局为分支结构的布局引导解析器。在区域充分性假设下，推导出前缀条件因子分解，使布局流与区域内容分支并行推进，解码深度降至最长布局-内容路径。该因子分解通过单个多模态大模型实现：打包可变长度祖先注意力在标准下一个词训练下保持可见性，掩码并行解码生成分支，由评估的vLLM后端作为并发请求处理，并复用缓存共享前缀。在OmniDocBench Full上，PaDoc取得91.1的总体布局F1，端到端解析器中最高综合得分94.24，以及最佳文本编辑距离0.038和公式CDM 95.59。在384页子集及单张A800 GPU上，其在五种并发级别下均为最快端到端解析器，有效页吞吐量提升67.4–118%，P95延迟降低39.2–54.9%（相比同主干序列微调基线）。代码已开源。
原文摘要 · Abstract (English)
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 parsers expose region-level parallelism at the cost of repeated visual prefills and fragmented page context. To retain full-page context while removing dependencies, we propose PaDoc, a layout-grounded parser that treats the predicted layout as a branching structure over a shared page representation. Under a region-sufficiency assumption, we derive a prefix-conditioned factorization in which the layout stream and regional content branches advance concurrently, reducing the decoding depth to the longest layout-content path. We realize this factorization within a single MLLM: packed variable-length ancestor attention preserves the visibility under standard next-token training, while masked parallel decoding creates branches that the evaluated vLLM backend serves as concurrent requests with cache-resident shared-prefix reuse. On OmniDocBench Full, PaDoc attains an Overall layout F1 of 91.1 and, among end-to-end parsers, a top-tier Overall score of 94.24 together with the best Text Edit (0.038) and Formula CDM (95.59). On a 384-page subset and one A800 GPU, it is the fastest end-to-end parser at five concurrency levels, improving valid-page throughput by 67.4-118% and reducing P95 latency by 39.2-54.9% relative to a same-backbone Sequential SFT baseline. Code is available at https://github.com/Longin-Yu/Padoc
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。