PDEBench and le_pde
About PDEBench
pdebench/PDEBench
PDEBench: An Extensive Benchmark for Scientific Machine Learning
This project provides a comprehensive benchmark for evaluating machine learning models designed to solve Partial Differential Equations (PDEs). It offers a wide range of realistic physical problems, along with ready-to-use datasets containing various initial/boundary conditions and PDE parameters. Scientists, engineers, and researchers working with scientific machine learning can use this to compare and develop methods for simulating complex physical phenomena.
About le_pde
snap-stanford/le_pde
LE-PDE accelerates PDEs' forward simulation and inverse optimization via latent global evolution, achieving significant speedup with SOTA accuracy
This project helps scientists and engineers quickly simulate complex physical systems and optimize designs. It takes in real-world data about a physical system, like weather patterns or material properties, and outputs rapid predictions for how that system will evolve over time, or suggests optimal configurations. Users include those working in weather forecasting, material science, and engine design who need to run many simulations quickly.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work