Physics Informed Fourier Neural Operator ( $\pi$ -FNO) is a physics-informed variant of regular FNO model, trained using physics constrained loss function. We show that $\pi$ -FNO can learn the weak solutions of nonlinear hyperbolic partial differential equations, which develop discontinuities even for smooth initial condition. A sample result is shown below. More details will be uploaded soon.
