NumpyTensorSpace.shape

property NumpyTensorSpace.shape

Number of scalar elements per axis.

Note

If dtype has a shape, we add it to the left of the given shape in the class creation. This is in contrast to NumPy, which adds extra axes to the right. We do this since we usually want to represent discretizations of vector- or tensor-valued functions by this, i.e., if dtype.shape == (3,) we expect f[0] to have shape shape.