Reproducible material for GaborPINN: Efficient physics informed neural networks using multiplicative filtered networks - Xinquan Huang, Tariq Alkhalifah.

Project structure

This repository is organized as follows:

• 📂 pinngabor: python library containing the main code and utils;
• 📂 asset: folder containing logo;
• 📂 data: folder containing data;
• 📂 scripts: set of python scripts used to run multiple experiments.

Getting started 👾 🤖

To ensure reproducibility of the results, we suggest using the pinnhash.yml file when creating an environment.

Simply run:

./install_env.sh

It will take some time, if at the end you see the word Done! on your terminal you are ready to go.

Remember to always activate the environment by typing:

conda activate pinnhash

Scripts

Go to folder scripts and run

bash run.sh

After running, go to folder exp/results/tb in the root_path produced by the procedures, and you could use tensorboard to visualize the trainig process and predictions.

Change the root_path

In the run.sh script, you need to modify the variable from line 4 to 6 (tb_root, run_root, data_root) to specify the root path for your procedures.

Check the results

After finish the training, you could go to the <run_root>/results/tb to use tensorboard --logdir=./ to check the training metrics and testing results.

Disclaimer: All experiments have been carried on a Intel(R) Xeon(R) CPU @ 2.10GHz equipped with a single NVIDIA GEForce A6000 GPU. Different environment configurations may be required for different combinations of workstation and GPU.

Cite us

@article{huang2023gaborpinn,
  title={GaborPINN: Efficient physics informed neural networks using multiplicative filtered},
  author={Huang, Xinquan and Alkhalifah, Tariq},
  journal={IEEE Geoscience and Remote Sensing Letters},
  volume={20},
  pages={1--5},
  year={2023},
  doi={10.1109/LGRS.2023.3330774},
  publisher={IEEE}
}

Acknowledgement

This code is developed based on open-sourced projects Multiplicative-filter-networks.