ifftn_op

→TensorImplPtr

int ifftn_op(const int & a, const int & n, const int & axes)

source edit

Python wrappers

pylucid.fft.ifftnfunction

N-dimensional inverse complex-to-complex DFT (forward of the inverse).

Inverts an N-D DFT. Uses the "backward" normalisation convention: forward has no scaling, this op divides by $\prod_j N_j$ where $N_j$ is the transform length along each axis.

Math

x[n_1, \dots, n_d] = \frac{1}{\prod_j N_j} \sum_{k_1=0}^{N_1-1} \cdots \sum_{k_d=0}^{N_d-1} X[k_1, \dots, k_d] \, \exp\!\left( +2\pi i \sum_{j=1}^{d} \frac{k_j n_j}{N_j} \right) .

See Also

fftn_op : Forward N-D DFT. irfftn_op : Inverse of the real-input DFT.

Parameters

aTensorImplPtr

Input tensor. C64 expected; F16/F32 are accepted and promoted.

nvector<int64_t>

Per-axis transform length. Empty means "use input sizes". When supplied, len(n) == len(axes) is required.

axesvector<int>

Axes to transform. Negative values are normalised against the input rank. Empty means "transform all axes".

Returns

TensorImplPtr

C64 tensor of the same shape as a (with axes[i] replaced by n[i] when n is supplied).

Raises

LucidError

If a is null, rank-0, has an unsupported dtype, or if len(n) != len(axes) when n is non-empty.

Notes

FFT bypasses IBackend dispatch and calls MLX directly from both CPU and GPU device paths (see _Detail.h).