不训练模型，通过预测中间状态加速分子结构生成。
Predictive Feature Caching for Training-free Acceleration of Molecular Geometry Generation
- 直接在等变网络上预测求解器各步的隐藏状态，实现推理加速。
- 在GEOM-Drugs数据集上，生成速度提升2倍，最高达3倍，质量几乎不变。
- 兼容预训练模型，可与多种优化方法叠加，最高提速7倍。
流匹配模型能生成高保真分子结构，但推理时需数百次网络计算，成为大规模采样时的主要瓶颈。本文提出一种无需训练的特征缓存策略，通过预测求解器各步骤的中间隐藏状态来加速生成。该方法作用于SE(3)-等变主干网络，兼容预训练模型，且与现有基于训练的加速和系统级优化正交。在GEOM-Drugs数据集上的实验表明，缓存使推理时间减少一半（墙钟时间），在保持样本质量的前提下速度提升高达3倍；当与其他通用、无损优化结合时，总加速比可达7倍。
原文摘要 · Abstract (English)
Flow matching models generate high-fidelity molecular geometries but incur significant computational costs during inference, requiring hundreds of network evaluations. This inference overhead becomes the primary bottleneck when such models are employed in practice to sample large numbers of molecular candidates. This work discusses a training-free caching strategy that accelerates molecular geometry generation by predicting intermediate hidden states across solver steps. The proposed method operates directly on the SE(3)-equivariant backbone, is compatible with pretrained models, and is orthogonal to existing training-based accelerations and system-level optimizations. Experiments on the GEOM-Drugs dataset demonstrate that caching achieves a twofold reduction in wall-clock inference time at matched sample quality and a speedup of up to 3x compared to the base model with minimal sample quality degradation. Because these gains compound with other optimizations, applying caching alongside other general, lossless optimizations yield as much as a 7x speedup.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。