Ian Williamson

Looks like that directory is missing from the pip package for some reason. You can just download the code from this repo instead.

Thanks for your interest in this! We are aware of that paper, but unfortunately we can't apply the scheme they propose here because the wave equation with loss (from the...

I'm definitely interested to learn about your project and what you hope to do. We would certainly be open to collaboration if there's an opportunity.

This is now partially implemented. Currently, the individual time step is a primitive. This seems to help with memory utilization during training, especially with nonlinearity. Perhaps we could investigate if...

Note that this is low priority to fix, just wanted to post it here to keep track of things.

Noting from our discussion earlier: this appears to be from the polar form of the the derivative whenever r=0.

Can you please share your code? It is difficult to know what could be going on without seeing it. Also, please do not post screenshots of code into issues /...

@RicardoSendrea07 I think you are misunderstanding the simulation and the physics. Why do you expect the mode overlap value to change with the position of the probe? The waveguide is...

The overlap value is going to be proportional to the amplitude of the eigenmode propagating through that cross section of the simulation. I say _proportional to_ because the function, as...

@RicardoSendrea07 From looking at your field plot, it appears that your eigenmode source is diffracting a bit (energy propagating obliquely up and down from the waveguide). This indicates that your...