search

tum-pbs / PhiFlow

A differentiable PDE solving framework for machine learning
MIT License
1.43k stars 193 forks source link

Fourier neural operators (PyTorch) #83

Closed kbali1297 closed 1 year ago

kbali1297 commented 1 year ago

Fourier Neural Operators for Pytorch. Single Operator consists of 4 Fourier layers (each layer comprising of SpectralConv and Conv layers). The overall network:

  1. Lift the input to the desire channel dimension by self.fc0 linear layer.
  2. 4 layers of the integral operators u' = (W + K)(u). W defined by self.w; K defined by self.conv.
  3. Project from the channel space to the output space by self.fc1 and self.fc2 linear layers.