Diffusion models have become widely popular for automated floorplan generation, producing highly realistic layouts conditioned on user-defined constraints. However, optimizing for perceptual metrics such as the Fréchet Inception Distance (FID) causes limited design diversity. To address this, we propose the Diversity Score (DS), a metric that quantifies layout diversity under fixed constraints. Moreover, to improve geometric consistency, we introduce a Boundary Cross-Attention (BCA) module that enables conditioning on building boundaries. Our experiments show that BCA significantly improves boundary adherence, while prolonged training drives diversity collapse undiagnosed by FID, revealing a critical trade-off between realism and diversity. Out-Of-Distribution evaluations further demonstrate the models' reliance on dataset priors, emphasizing the need for generative systems that explicitly balance fidelity, diversity, and generalization in architectural design tasks.

翻译：扩散模型在自动化平面图生成中已得到广泛应用，能够根据用户定义的约束条件生成高度真实的布局。然而，针对感知指标（如Fréchet Inception Distance，FID）的优化会导致设计多样性受限。为解决这一问题，我们提出了多样性评分（Diversity Score，DS），这是一种在固定约束条件下量化布局多样性的指标。此外，为提高几何一致性，我们引入了边界交叉注意力（Boundary Cross-Attention，BCA）模块，该模块能够基于建筑边界进行条件生成。实验表明，BCA显著提升了边界贴合度，而长时间训练则会导致FID无法诊断的多样性崩溃，揭示了真实性与多样性之间的关键权衡。分布外评估进一步证明了模型对数据集先验的依赖，强调了在建筑设计任务中需要构建能够明确平衡保真度、多样性与泛化能力的生成系统。