Fast3Dcache：免训练的3D几何合成加速
Fast3Dcache: Training-free 3D Geometry Synthesis Acceleration
提出一种无需训练的几何感知缓存框架，通过预测性缓存调度约束和时空稳定性准则，在加速3D扩散模型推理的同时保持几何保真度，可复用于其他3D生成任务的加速。
原文依据 we propose Fast3Dcache, a training-free geometry-aware caching framework that accelerates 3D diffusion inference while preserving geometric fidelity.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论扩散模型在二维图像、视频和三维形状等模态上已取得令人印象深刻的生成质量，但由于迭代去噪过程，其推理仍然计算昂贵。尽管近期基于缓存的方法通过重用冗余计算有效加速了二维和视频生成，但直接将这些技术应用于三维扩散模型会严重破坏几何一致性。在三维合成中，缓存潜变量特征中的微小数值误差会累积，导致结构伪影和拓扑不一致。为克服这一局限，我们提出Fast3Dcache，一种免训练的几何感知缓存框架，在保持几何保真度的同时加速三维扩散推理。我们的方法引入预测性缓存调度约束（PCSC），根据体素稳定模式动态确定缓存配额，并引入时空稳定性准则（SSC），基于速度幅值和加速度准则选择稳定特征进行重用。综合实验表明，Fast3Dcache显著加速推理，实现高达27.12%的加速和54.83%的FLOPs降低，同时几何质量下降极小，以倒角距离（2.48%）和F分数（1.95%）衡量。
英文原文摘要 1299 字符
Diffusion models have achieved impressive generative quality across modalities like 2D images, videos, and 3D shapes, but their inference remains computationally expensive due to the iterative denoising process. While recent caching-based methods effectively reuse redundant computations to speed up 2D and video generation, directly applying these techniques to 3D diffusion models can severely disrupt geometric consistency. In 3D synthesis, even minor numerical errors in cached latent features accumulate, causing structural artifacts and topological inconsistencies. To overcome this limitation, we propose Fast3Dcache, a training-free geometry-aware caching framework that accelerates 3D diffusion inference while preserving geometric fidelity. Our method introduces a Predictive Caching Scheduler Constraint (PCSC) to dynamically determine cache quotas according to voxel stabilization patterns and a Spatiotemporal Stability Criterion (SSC) to select stable features for reuse based on velocity magnitude and acceleration criterion. Comprehensive experiments show that Fast3Dcache accelerates inference significantly, achieving up to a 27.12% speed-up and a 54.83% reduction in FLOPs, with minimal degradation in geometric quality as measured by Chamfer Distance (2.48%) and F-Score (1.95%).