Kernel size of PrmEmbd module

[yhy258]

Thanks for your code sharing.

CAPE-ML4Sci/CAPE/models/PrmEmbd/PrmEmbd.py

Lines 293 to 305 in 8ae7d64

	class PrmEmb_Block_2d(nn.Module):
	"""A 2-layer parameter embedding module for 2D data."""
	def __init__(self,
	widening_factor: int = 16,
	kernel_size: int = 5,
	num_params: int = 1,
	if_11cnv = False,
	num_channels: int = 1,
	num_channels_PrmEmb: int = 1,
	modes = 9,
	normed_dim=[64,64]
	):

According to this code, I think the PrmEmbd's kernel size is set as 5. In this case, doesn't this setting violate the discretization invariance properties of Operator models?