/NeuralOperator

Primary LanguagePython

NeuralOperator

This repository explores the neural operator approach for creating surrogates for PDE solvers.

Three steps to test GalerkinNO:

  1. Create a folder 'data',which has two subfolders 'burgers_1d' and 'darcy_2d' containing the data files.

  2. Modify configs in the file 'config_1D.yml' or 'config_2D.yml'.

  3. Run 'FFT_1D.py' or 'FFT_2D.py'.

References

  1. Li, Zongyi, Nikola Kovachki, Kamyar Azizzadenesheli, Burigede Liu, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).

  2. Li, Zongyi, Daniel Zhengyu Huang, Burigede Liu, and Anima Anandkumar. "Fourier neural operator with learned deformations for pdes on general geometries." Journal of Machine Learning Research 24, no. 388 (2023): 1-26.

Some tips:

  1. Request a peer review
    • Modify the code
    • Start a new branch git checkout -b "your-branch-name". Then you will be in your branch.
    • Modify the code, commit and push ....... When you first push, it might give errors, try git push --set-upstream origin your-branch-name
    • When you are ready to merge, push the code. Then go to the GitHub webpage, switch to your branch, click open pull request. You will create a pull request. Describe your change and assign reviewers.
    • When reviewers agree with your changes, you can push the code