Published May 2023 | Version v2
Dataset Open

Representative PDE Benchmarks (RPB) - Convolutional Neural Operators for robust and accurate learning of PDEs

  • 1. ROR icon ETH Zurich

Description

Representative PDE Benchmarks (RPB)

We made the datasets as clean as we could. It is now clear which datasets correspong to in-distribution and out-of-distributions testings. Please read our BLOG!

There are 8 benchmarks that we considered:

  • Poisson Equation
  • Wave Equation
  • Smooth Transport
  • Discontinuous Transport
  • Allen-Cahn Equation
  • Navier-Stokes Equations
  • Darcy Flow
  • Compressible Euler Equations

REMARK: DO NOT USE __MACOSX FILES, BUT THE .H5 ONES! 

 

Please refer to out github page to get the script that loads the datasets: 

https://github.com/bogdanraonic3/ConvolutionalNeuralOperator/tree/main 

 

Please let us know if there are mistakes in the data (email: braonic@student.ethz.ch)

Files

data.zip

Files (2.4 GB)

Name Size Download all
md5:268a51a9c90f4ee01cd0c1495664ec25
2.4 GB Preview Download