Generative models for crystalline materials often rely on equivariant graph neural networks, which capture geometric structure well but are costly to train and slow to sample. We present Crystalite, a lightweight diffusion Transformer for crystal modeling built around two simple inductive biases. The first is Subatomic Tokenization, a compact chemically structured atom representation that replaces high-dimensional one-hot encodings and is better suited to continuous diffusion. The second is the Geometry Enhancement Module (GEM), which injects periodic minimum-image pair geometry directly into attention through additive geometric biases. Together, these components preserve the simplicity and efficiency of a standard Transformer while making it better matched to the structure of crystalline materials. Crystalite achieves state-of-the-art results on crystal structure prediction benchmarks, and de novo generation performance, attaining the best S.U.N. discovery score among the evaluated baselines while sampling substantially faster than geometry-heavy alternatives.

翻译：晶体材料的生成模型通常依赖等变图神经网络，这类网络虽能有效捕捉几何结构，但训练成本高且采样速度慢。我们提出Crystalite——一种基于两种简单归纳偏置构建的轻量级扩散Transformer晶体建模方法。第一种是亚原子分词化（Subatomic Tokenization），一种紧凑的化学结构化原子表示，替代了高维独热编码，更适合连续扩散过程。第二种是几何增强模块（Geometry Enhancement Module, GEM），通过加性几何偏置将周期性最小成像对几何直接注入注意力机制。这些组件共同保留了标准Transformer的简洁性与高效性，同时使其更适配晶体材料的结构特性。Crystalite在晶体结构预测基准和从头生成任务中均达到最优性能，在评估基线中取得最佳S.U.N.发现分数，且采样速度显著快于几何密集型替代方法。