Skip to Content

fftn_op

View the code on GitHub

Structs

Struct: FFTN

Fields

Methods

fwd(arg0: Array, dims: List[Int], norm: String) -> Array
Sets up the Array object for the FFT operation.
Args
  • arg0: Array

  • dims: List[Int]

  • norm: String

Returns
  • Array
jvp(primals: List[Array], tangents: List[Array]) -> Array
Computes the Jacobian-vector product for the FFT function.
Args
  • primals: List[Array]

  • tangents: List[Array]

Returns
  • Array
vjp(primals: List[Array], grad: Array, out: Array) -> List[Array]
Computes the vector-Jacobian product for the FFT function.
Args
  • primals: List[Array]

  • grad: Array

  • out: Array

Returns
  • List[Array]
__call__(mut curr: Array, args: List[Array])
Executes the FFT operation inplace.
Args
  • curr: Array

  • args: List[Array]

Functions

fftn

fftn(x: Array, dims: List[Int] = List(), norm: String = String("backward")) -> Array
Compute the n-dimensional FFT.
Args
  • x: Array The input array.

  • dims: List[Int] (default: List()) The dimensions along which to compute the FFT.

  • norm: String (default: String("backward")) The normalization mode.

Returns
  • Array - The n-dimensional FFT of the input array.
Last updated on