Infinity-Parser2 技术报告

arXiv:2607.07836v1 公告类型：新 摘要：我们提出了 Infinity-Parser2，这是一种大型多模态模型，它将可控数据合成管道与用于端到端文档解析的多任务强化学习相结合，解决了忠实注释的解析语料库的持续稀缺问题。我们的贡献是三重的。首先，我们构建了一个可扩展的合成引擎，将可控渲染框架与迭代细化循环配对，并用它构建并开源了 Infinity-Doc2-5M：一个包含 500 万个样本的双语（中文/英文）语料库，涵盖不同的文档类型，用元素边界框、规范内容形式（Markdown、HTML、LaTeX、SMILES、结构化图表）和整页阅读顺序进行注释。 Second, we introduce a verifiable, multi-task reward system that enables Joint Reinforcement Learning across eight co-trained objectives (document parsing, layout analysis, table parsing, math formula parsing, chart parsing, chemical formula parsing, document VQA, and general multimodal understanding), unifying perception, structure, and reasoning in a single optimization signal. Third, we release two variants under a shared architecture: Infinity-Parser2-Flash, optimized for low-latency inference with a $3.68\times$ throughput gain over Infinity-Parser-7B, and Infinity-Parser2-Pro, engineered for precision-critical settings. Infinity-Parser2-Pro reaches state-of-the-art 87.6% on olmOCR-Bench and 74.3% on ParseBench, surpassing DeepSeek-OCR-2, PaddleOCR-VL-1.5, and MinerU2.5, with strong generalization to charts, chemical formulas, and document VQA.