pihnn.nn.ComplexParameter#

class pihnn.nn.ComplexParameter(minvalue=-1 - 1j, maxvalue=1 + 1j, in_features=1, out_features=1)#

Bases: torch.nn.Module

Extension of torch.nn.Parameter to complex values.

Parameters:
  • minvalue (complex) – The minimum value in the parameter initialization. Specifically, parameters are initialized to \(p \sim \mathcal{U}(-\texttt{minvalue},\texttt{maxvalue})\).

  • maxvalue (complex) – The minimum value in the parameter initialization. See above.

  • in_features (int) – First size of parameter tensor.

  • out_features (int) – Second size of parameter tensor.

forward()#

It returns the parameter tensor.

Returns:

parameter (torch.tensor) - The tensor of size (in_features,out_features) with the parameters.