magic_intpl

class diffsptk.MagicNumberInterpolation(magic_number=0)

See this page for details.

Parameters:

magic_numberfloat

The magic number to be interpolated.

forward(x)

Interpolate magic number.

Parameters:

xTensor [shape=(B, N, D) or (N, D) or (N,)]

The data containing magic number.

Returns:

outTensor [shape=(B, N, D) or (N, D) or (N,)]

The data after interpolation.

Examples

>>> x = torch.tensor([0, 1, 2, 0, 4, 0]).float()
>>> x
tensor([0., 1., 2., 0., 4., 0.])
>>> magic_intpl = diffsptk.MagicNumberInterpolation(0)
>>> y = magic_intpl(x)
>>> y
tensor([1., 1., 2., 3., 4., 4.])

diffsptk.functional.magic_intpl(x, magic_number=0)

Interpolate magic number.

Parameters:

xTensor [shape=(B, N, D) or (N, D) or (N,)]

The data containing magic number.

magic_numberfloat or Tensor

The magic number to be interpolated.

Returns:

outTensor [shape=(B, N, D) or (N, D) or (N,)]

The data after interpolation.

See also

pitch