Smaller But Better: Unifying Layout Generation with Smaller Large Language Models

TL;DR

This paper introduces LGGPT, a compact 1.5B parameter language model that unifies layout generation tasks using a novel instruction-response format and encoding strategy, achieving high performance efficiently.

Contribution

The paper presents ALI and ULR as a unified I/O template and IQE for efficient layout encoding, enabling a small LLM to excel in diverse layout generation tasks.

Findings

LGGPT outperforms larger models in unified layout generation tasks.

The proposed encoding and template improve efficiency and performance.

A 1.5B model balances proficiency and efficiency effectively.

Abstract

We propose LGGPT, an LLM-based model tailored for unified layout generation. First, we propose Arbitrary Layout Instruction (ALI) and Universal Layout Response (ULR) as the uniform I/O template. ALI accommodates arbitrary layout generation task inputs across multiple layout domains, enabling LGGPT to unify both task-generic and domain-generic layout generation hitherto unexplored. Collectively, ALI and ULR boast a succinct structure that forgoes superfluous tokens typically found in existing HTML-based formats, facilitating efficient instruction tuning and boosting unified generation performance. In addition, we propose an Interval Quantization Encoding (IQE) strategy that compresses ALI into a more condensed structure. IQE precisely preserves valid layout clues while eliminating the less informative placeholders, facilitating LGGPT to capture complex and variable layout generation conditions during the unified training process. Experimental results demonstrate that LGGPT achieves superior or on par performance compared to existing methods. Notably, LGGPT strikes a prominent balance between proficiency and efficiency with a compact 1.5B parameter LLM, which beats prior 7B or 175B models even in the most extensive and challenging unified scenario. Furthermore, we underscore the necessity of employing LLMs for unified layout generation and suggest that 1.5B could be an optimal parameter size by comparing LLMs of varying scales. Code is available at https://github.com/NiceRingNode/LGGPT.