Visualize the points that are used for training in NeuralPDE

Specific Domains Machine Learning
, , ,
1

Hi,

I would like to print and visualize the points that are being selected at each iteration by the QuasiRandomTraining strategy in NeuralPDE.

I have checked the many fields of symbolic_discretize but I cannot pinpoint which of them returns the points used.

Is there any way to do that or is this functionality not supported?

Thanks in advance for the help.

2

Currently it’s not supported because it resamples every time and those samples are not saved. It would actually be very data intensive to save this because if you do 1000 iterations for 1000 samples you’d have a (1000,1000,d) dataset of samples… which is certainly not something you’d want to store for big d if it could even fit in your RAM.