Youtu-Parsing: Perception, Structuring and Recognition via High-Parallelism Decoding

TL;DR

Youtu-Parsing introduces a high-parallelism decoding framework using Vision Transformers and prompt-guided language models, achieving state-of-the-art document parsing speed and accuracy across diverse document elements and challenging content types.

Contribution

The paper presents a novel high-parallelism decoding strategy for document parsing, combining token and query parallelism with a versatile architecture for improved speed and robustness.

Findings

5-11x faster decoding than traditional methods

State-of-the-art performance on OmniDocBench and olmOCR-bench

Robust handling of multilingual, handwritten, and rare characters

Abstract

This paper presents Youtu-Parsing, an efficient and versatile document parsing model designed for high-performance content extraction. The architecture employs a native Vision Transformer (ViT) featuring a dynamic-resolution visual encoder to extract shared document features, coupled with a prompt-guided Youtu-LLM-2B language model for layout analysis and region-prompted decoding. Leveraging this decoupled and feature-reusable framework, we introduce a high-parallelism decoding strategy comprising two core components: token parallelism and query parallelism. The token parallelism strategy concurrently generates up to 64 candidate tokens per inference step, which are subsequently validated through a verification mechanism. This approach yields a 5--11x speedup over traditional autoregressive decoding and is particularly well-suited for highly structured scenarios, such as table recognition. To further exploit the advantages of region-prompted decoding, the query parallelism strategy enables simultaneous content prediction for multiple bounding boxes (up to five), providing an additional 2x acceleration while maintaining output quality equivalent to standard decoding. Youtu-Parsing encompasses a diverse range of document elements, including text, formulas, tables, charts, seals, and hierarchical structures. Furthermore, the model exhibits strong robustness when handling rare characters, multilingual text, and handwritten content. Extensive evaluations demonstrate that Youtu-Parsing achieves state-of-the-art (SOTA) performance on both the OmniDocBench and olmOCR-bench benchmarks. Overall, Youtu-Parsing demonstrates significant experimental value and practical utility for large-scale document intelligence applications.