OmniDocLayout: Towards Diverse Document Layout Generation via Coarse-to-Fine LLM Learning

TL;DR

This paper introduces OmniDocLayout, a large-scale dataset and a novel two-stage LLM approach for diverse and complex document layout generation, significantly advancing the field beyond traditional Manhattan-style structures.

Contribution

It presents the first million-scale dataset of diverse document layouts and a two-stage Coarse-to-Fine LLM training paradigm for improved layout generation across multiple domains.

Findings

Outperforms existing layout generation methods and general-purpose LLMs

Achieves strong results on multiple domain datasets

Substantially improves diversity and complexity in generated layouts

Abstract

Document AI has advanced rapidly and is attracting increasing attention. Yet, while most efforts have focused on document layout analysis (DLA), its generative counterpart, layout generation, remains underexplored. Distinct from traditional graphic layout design and room layout planning, document layout generation typically involves a larger number of elements per page and exhibits greater structural diversity and complexity. Currently, a major obstacle lies in the scarcity of diverse document layouts: academic papers with Manhattan-style structures dominate existing studies, while open-world genres such as newspapers and magazines remain severely underrepresented. To address this gap, we curate OmniDocLayout-1M, the first million-scale dataset of diverse document layouts, covering six common document types and comprising contemporary layouts collected from multiple sources. Moreover, since existing methods struggle in complex domains and often fail to arrange long sequences coherently, we introduce OmniDocLayout-LLM, a 0.5B model with designed two-stage Coarse-to-Fine learning paradigm:1) learning universal layout principles from our dataset with coarse category definitions, and 2) transferring the knowledge to a specific domain with few fine-grained annotated samples. Extensive experiments demonstrate that our approach achieves strong performance on multiple domains in M$^6$Doc dataset, substantially surpassing both existing layout generation experts and several latest general-purpose LLMs. Our code, dataset, and models will be publicly released.