Transolver-3:将Transformer求解器扩展到工业级几何规模
Transolver-3: Scaling Up Transformer Solvers to Industrial-Scale Geometries
AI总结:
Transolver-3通过矩阵乘法结合性质和几何切片技术,实现高保真物理模拟,可处理超1.6亿单元的工业级网格,通过随机子集训练和物理状态缓存提升性能。
摘要:
深度学习已涌现出作为神经替代模型求解偏微分方程(PDEs)的 transformative 工具,即神经 PDE 求解器。然而,将这些求解器扩展到具有超过 $10^8 $ 个单元的工业级几何结构仍是一个根本性挑战,因为处理高分辨率网格的内存复杂性是 prohibitive 的。我们提出了 Transolver-3,作为 Transolver 家族的新成员,是一个高度可扩展的框架,用于高保真物理模拟。为了弥合有限 GPU 容量与复杂工程任务的分辨率要求之间的差距,我们引入了两种关键的架构优化:通过利用矩阵乘法结合性质和几何切片技术加速切片和去切片,并通过几何切片瓷砖化来划分物理状态的计算。结合通过在随机子集上学习原始高分辨率网格的 amortized 训练策略和推理过程中的物理状态缓存技术,Transolver-3 能够在工业级网格上实现高保真场预测。广泛的实验表明,Transolver-3 可以处理超过 1.6 亿个单元的网格,在三个具有挑战性的模拟基准测试中取得了令人印象深刻的表现,包括飞机和汽车设计任务。代码可在 https://github.com/thuml/Transolver-3 上获得。
英文摘要:
Deep learning has emerged as a transformative tool for the neural surrogate modeling of partial differential equations (PDEs), known as neural PDE solvers. However, scaling these solvers to industrial-scale geometries with over $10^8$ cells remains a fundamental challenge due to the prohibitive memory complexity of processing high-resolution meshes. We present Transolver-3, a new member of the Transolver family as a highly scalable framework designed for high-fidelity physics simulations. To bridge the gap between limited GPU capacity and the resolution requirements of complex engineering tasks, we introduce two key architectural optimizations: faster slice and deslice by exploiting matrix multiplication associative property and geometry slice tiling to partition the computation of physical states. Combined with an amortized training strategy by learning on random subsets of original high-resolution meshes and a physical state caching technique during inference, Transolver-3 enables high-fidelity field prediction on industrial-scale meshes. Extensive experiments demonstrate that Transolver-3 can handle meshes with over 160 million cells, achieving impressive performance across three challenging simulation benchmarks, including aircraft and automotive design tasks. Code is available at https://github.com/thuml/Transolver-3.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。