fn
histogramdd
→tuplehistogramdd(input: Tensor, bins: int | Sequence[int] = 10, range: Sequence[tuple[float, float]] | None = None, density: bool = False)Implementing kernel
C++ engine symbols that back this Python API.N-dimensional histogram for arbitrary sample dimension.
Parameters
inputTensorShape
(N, D) — samples in
dimensions. Other ranks raise.binsint or sequence of int= 10Bin count per axis. A single int applies to every axis; a
sequence of length sets each axis independently.
Default
10.rangesequence of (float, float)= NonePer-axis
(lo, hi) ranges. When None (default) each
axis uses its own (min, max) of the samples.densitybool= FalseWhen
True divide by total count × bin volume to yield a
density that integrates to 1. Default False.Returns
tuple(counts, edges). counts shape
(bins_0, …, bins_{D-1}); edges is a length-
list of 1-D tensors of size bins_i + 1.