Utility Functions

MeshGraphNets.data_minmaxFunction
data_minmax(path)

Calculates the minimum and maximum of each numeric feature across the training, validation, and test splits. For target features, it also calculates the extrema of their finite-difference derivatives.

Arguments

  • path: Path to the dataset files.

Returns

  • Dictionary mapping feature names to [minimum, maximum]. Derivative statistics use keys of the form "target|<feature>".
source
MeshGraphNets.data_meanstdFunction
data_meanstd(path)

Calculates the mean and standard deviation of each numeric feature across the training, validation, and test splits. For target features, it also calculates statistics for their finite-difference derivatives.

Arguments

  • path: Path to the dataset files.

Returns

  • Dictionary mapping feature names to (mean, standard deviation). Derivative statistics use keys of the form "target|<feature>".
source