This folder contains files of python code accompanying the manuscript "Gradient statistics based multi-objective optimization in Physics Informed Neural Networks"

The folders contains 4 python files:

pinn.py ----> contains functions to create Physics Informed Neural Networks

grad_stats.py ---> contains functions to calculate gradient statistics mentioned in the paper

PoissonEqn.ipynb ----> Jupyter Notebook for solving Poissons Equation using all the weighting schemes.

KleinGordonEqn.ipynb----> Jupyter Notebook for solving Klein-Gordon Equations using all the weighting schemes.

To run the notebooks:

1. Please download all the files into a single folder.
2. Prerequisite libraries are: Pytorch, Numpy, Matplotlib, Tqdm.
3. Run the cells to recreate graphs.