PINNs

Framework for physics-informed neural networks.

Implementation for Helmholtz equations and Navier-Lame equations. Easy extendable to solve any PDE with physics-informed neural networks.

Several classes included to easily use Tensorboard or RayTune in combination with PINN training.

Trained models and reference FEM data for the sample scripts available upon request.

Following scripts work without the need of any additional dataset:

• main_1Dhelmholtz_equation.py
• main_2Dhelmholtz_equation.py