cupynumeric.convolve#
- cupynumeric.convolve( ) ndarray#
Returns the discrete, linear convolution of two ndarrays.
If a and v are both 1-D and v is longer than a, the two are swapped before computation. For N-D cases, the arguments are never swapped.
- Parameters:
a ((N,) array_like) – First input ndarray.
v ((M,) array_like) – Second input ndarray.
mode (
{'full', 'valid', 'same'}, optional) –- ‘same’:
The output is the same size as a, centered with respect to the ‘full’ output. (default)
- ’full’:
The output is the full discrete linear convolution of the inputs.
- ’valid’:
The output consists only of those elements that do not rely on the zero-padding. In ‘valid’ mode, either a or v must be at least as large as the other in every dimension.
method (
{'auto', 'direct', 'fft'}, optional) –A string indicating which method to use to calculate the convolution.
- ’auto’:
Automatically chooses direct or Fourier method based on an estimate of which is faster (default)
- ’direct’:
The convolution is determined directly from sums, the definition of convolution
- ’fft’:
The Fourier Transform is used to perform the convolution
- Returns:
out – Discrete, linear convolution of a and v.
- Return type:
See also
Notes
The current implementation only supports the ‘same’ mode.
Unlike numpy.convolve, cupynumeric.convolve supports N-dimensional inputs, but it follows NumPy’s behavior for 1-D inputs.
- Availability:
Multiple GPUs, Multiple CPUs