Efficient Document Parsing via Parallel Token Prediction
On the map: Multimodal LLMs / vision-language models → Document understanding and multimodal RAG
Abstract (arXiv)
Document parsing, as a fundamental yet crucial vision task, is being revolutionized by vision-language models (VLMs). However, the autoregressive (AR) decoding inherent to VLMs creates a significant bottleneck, severely limiting parsing speed. In this paper, we propose Parallel-Token Prediction (PTP), a plugable, model-agnostic and simple-yet-effective method that enables VLMs to generate multiple future tokens in parallel with improved sample efficiency. Specifically, we insert some learnable tokens into the input sequence and design corresponding training objectives to equip the model with parallel decoding capabilities for document parsing. Furthermore, to support effective training, we develop a comprehensive data generation pipeline that efficiently produces large-scale, high-quality document parsing training data for VLMs. Extensive experiments on OmniDocBench and olmOCR-bench demonstrate that our method not only significantly improves decoding speed (1.6x-2.2x) but also reduces model hallucinations and exhibits strong generalization abilities.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Document understanding and multimodal RAG, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 37 of its 48 references that are in the map. First posted on arXiv 16 Mar 2026.
Builds on
37 of 48 references.
Closest ancestors · 3
Same direction, newest first
- 2025PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model16 Oct
- 2025MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing26 Sep
- 2025POINTS-Reader: Distillation-Free Adaptation of Vision-Language Models for Document Conversion1 Sep
Other references · 34
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
- 2023GPT-4 Technical Reportcited by 18,396
- 2025Qwen2.5-VL Technical Reportcited by 5,587
- 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
- 2022Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answeringcited by 1,919
- 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
- 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
- 2023Vision Transformers Need Registerscited by 871
- 2024Better & Faster Large Language Models via Multi-token Predictioncited by 360
- 2018Blockwise Parallel Decoding for Deep Autoregressive Modelscited by 323
- 2024MinerU: An Open-Source Solution for Precise Document Content Extractioncited by 267
- 2025DeepSeek-OCR: Contexts Optical Compressioncited by 183
22 more
- 2024General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Modelcited by 180
- 2024OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotationscited by 150
- 2022Unifying Vision, Text, and Layout for Universal Document Processingcited by 140
- 2025MiMo: Unlocking the Reasoning Potential of Language Model -- From Pretraining to Posttrainingcited by 121
- 2025MonkeyOCR: Document Parsing with a Structure-Recognition-Relation Triplet Paradigmcited by 87
- 2025Dolphin: Document Image Parsing via Heterogeneous Anchor Promptingcited by 79
- 2022Syntax-Aware Network for Handwritten Mathematical Expression Recognitioncited by 68
- 2025olmOCR 2: Unit Test Rewards for Document OCRcited by 47
- 2024Document Parsing Unveiled: Techniques, Challenges, and Prospects for Structured Information Extractioncited by 44
- 2024UniMERNet: A Universal Network for Real-World Mathematical Expression Recognitioncited by 42
- 2024Image Over Text: Transforming Formula Recognition Evaluation with Character Detection Matchingcited by 41
- 2024OCR Hinders RAG: Evaluating the Cascading Impact of OCR on Retrieval-Augmented Generationcited by 26
- 2025PP-DocLayout: A Unified Document Layout Detection Model to Accelerate Large-Scale Data Constructioncited by 26
- 2022Wukong-Reader: Multi-modal Pre-training for Fine-grained Visual Document Understandingcited by 20
- 2023Context Perception Parallel Decoder for Scene Text Recognitioncited by 19
- 2024Revolutionizing Retrieval-Augmented Generation with Enhanced PDF Structure Recognitioncited by 16
- 2025Multi-Token Prediction Needs Registerscited by 16
- 2025FastMTP: Accelerating LLM Inference with Enhanced Multi-Token Predictioncited by 13
- 2025Seeing is Believing? Mitigating OCR Hallucinations in Multimodal Large Language Modelscited by 12
- 2023A CTC Alignment-based Non-autoregressive Transformer for End-to-end Automatic Speech Recognitioncited by 5
- 2024Non-autoregressive Sequence-to-Sequence Vision-Language Modelscited by 5
- 2024DocFusion: A Unified Framework for Document Parsing Taskscited by 5
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2026HPD-Parsing: Hierarchical Parallel Document Parsing
- 2026P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scaling
- 2026HSD: Training-Free Acceleration for Document Parsing Vision-Language Models with Hierarchical Speculative Decoding
- 2025Parallel Token Prediction for Language Models
- 2026Towards Real-World Document Parsing via Realistic Scene Synthesis and Document-Aware Training
- 2026PaDoc: Layout-Grounded Parallel Decoding for Document Parsing
- 2026Dr. DocBench: A Comprehensive Benchmark for Expert-Level and Difficult Document Parsing
- 2025Index-Preserving Lightweight Token Pruning for Efficient Document Understanding in Vision-Language Models