Stockmark-Nemotron-3-Nano-Omni-JapanDocReader: Structured Document Parsing via Capability Injection and Forgetting Control

• URL: http://arxiv.org/abs/2608.06758v1
• Date: Fri, 07 Aug 2026 03:24:59 GMT
• Title: Stockmark-Nemotron-3-Nano-Omni-JapanDocReader: Structured Document Parsing via Capability Injection and Forgetting Control
• Abstract summary: We present a Japanese document understanding model built from Nemotron-3-Nano- Omni-30B-A3B-Reasoning-BF16.<n>We study parsing-centric SFT, which uses only structured document parsing data; mixed SFT, which combines structured document parsing and VQA data.<n>Applying DAPO-based parsing-centric RL on top of the mixed SFT checkpoint further improves structured document parsing beyond the SFT ceiling.
• Score: 2.661673377954457
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We present Stockmark-Nemotron-3-Nano-Omni-JapanDocReader, a Japanese document understanding model built from Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16. The central goal of this work is structured document parsing via capability injection and forgetting control: we inject Japanese structured document parsing capability into a reasoning-oriented multimodal model while preserving its document VQA capability as much as possible. We study parsing-centric SFT, which uses only structured document parsing data; mixed SFT, which combines structured document parsing and VQA data; and parsing-centric RL, which optimizes structured parsing with a task-level reward. Our experiments show that parsing-centric SFT substantially improves structured document parsing performance but causes measurable VQA forgetting. Mixed SFT mitigates this forgetting while preserving nearly the same structured parsing performance. Applying DAPO-based parsing-centric RL on top of the mixed SFT checkpoint further improves structured document parsing beyond the SFT ceiling, producing the final released model. The training data is constructed with a data engine consisting of two complementary synthetic streams: a Japanese Document VQA Stream and a programmatic structured document parsing stream. We also discuss reward design and variance-based prompt filtering for continuous structured document parsing rewards, highlighting their importance for making RL effective in long-reasoning structured document parsing tasks.

Related papers

• RT-DocLayout: Real-Time End-to-End Document Layout Analysis with Reading Order in the Wild [14.715243408844058]
  We present RT-Doc, a highly efficient end-to-end framework for document layout analysis.<n>The proposed model unifies classification, pixel-level segmentation, and geometric order reading order prediction for layout elements.<n> RT-Doc significantly improves full-document reconstruction quality, providing a scalable and practical foundation for real-world document intelligence systems.
  arXiv  Detail & Related papers  (2026-06-22T13:48:39Z)
• MPDocBench-Parse: Benchmarking Practical Multi-page Document Parsing [74.84107522458798]
  MPDocBench-Parse is a benchmark for multi-page document parsing in real-world applications.<n>It contains 433 manually annotated documents with 3,246 pages, covering 15 document types in English and Chinese, with diverse layout styles.
  arXiv  Detail & Related papers  (2026-05-21T07:36:41Z)
• MultiDocFusion: Hierarchical and Multimodal Chunking Pipeline for Enhanced RAG on Long Industrial Documents [30.113467423647865]
  RAG-based QA has emerged as a powerful method for processing long industrial documents.<n>We introduce MultiDocFusion, a multimodal chunking pipeline that integrates vision-based document parsing, OCR, text extraction and reconstruction.<n>Experiments demonstrate that MultiDocFusion improves retrieval precision by 8-15% and ANLS QA scores by 2-3% compared to baselines.
  arXiv  Detail & Related papers  (2026-04-14T06:40:22Z)
• Towards Real-World Document Parsing via Realistic Scene Synthesis and Document-Aware Training [23.87978106727431]
  We propose a data-training co-design framework for robust end-to-end document parsing.<n>Our method achieves superior accuracy and robustness across both scanned/digital and real-world captured scenarios.
  arXiv  Detail & Related papers  (2026-03-25T03:19:09Z)
• Training-Free Acceleration for Document Parsing Vision-Language Model with Hierarchical Speculative Decoding [102.88996030431662]
  We propose a training-free and highly efficient acceleration method for document parsing tasks.<n>Inspired by speculative decoding, we employ a lightweight document parsing pipeline as a draft model to predict batches of future tokens.<n>We demonstrate the effectiveness of our approach on the general-purpose OmniDocBench.
  arXiv  Detail & Related papers  (2026-02-13T14:22:10Z)
• MonkeyOCR v1.5 Technical Report: Unlocking Robust Document Parsing for Complex Patterns [80.05126590825121]
  MonkeyOCR v1.5 is a unified vision-language framework that enhances both layout understanding and content recognition.<n>To address complex table structures, we propose a visual consistency-based reinforcement learning scheme.<n>Two specialized modules, Image-Decoupled Table Parsing and Type-Guided Table Merging, are introduced to enable reliable parsing of tables.
  arXiv  Detail & Related papers  (2025-11-13T15:12:17Z)
• Infinity Parser: Layout Aware Reinforcement Learning for Scanned Document Parsing [46.14775667559124]
  layoutRL is an end-to-end reinforcement learning framework that trains models to be explicitly layout-aware.<n>We will publicly release our code and dataset to accelerate progress in robust document understanding.
  arXiv  Detail & Related papers  (2025-06-01T15:19:52Z)
• CREPE: Coordinate-Aware End-to-End Document Parser [13.530212337717515]
  We formulate an OCR-free sequence generation model for visual document understanding (VDU) Our model not only parses text from document images but also extracts the spatial coordinates of the text based on the multi-head architecture. Named as Coordinate-aware End-to-end Document. CREPE, our method uniquely integrates these capabilities by introducing a special token for OCR text.
  arXiv  Detail & Related papers  (2024-05-01T00:30:13Z)
• Document Structure in Long Document Transformers [64.76981299465885]
  Long documents often exhibit structure with hierarchically organized elements of different functions, such as section headers and paragraphs. Despite the omnipresence of document structure, its role in natural language processing (NLP) remains opaque. Do long-document Transformer models acquire an internal representation of document structure during pre-training? How can structural information be communicated to a model after pre-training, and how does it influence downstream performance?
  arXiv  Detail & Related papers  (2024-01-31T08:28:06Z)
• mPLUG-DocOwl: Modularized Multimodal Large Language Model for Document Understanding [55.4806974284156]
  Document understanding refers to automatically extract, analyze and comprehend information from digital documents, such as a web page. Existing Multi-model Large Language Models (MLLMs) have demonstrated promising zero-shot capabilities in shallow OCR-free text recognition.
  arXiv  Detail & Related papers  (2023-07-04T11:28:07Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.