Struct Shape 

pub struct Shape(/* private fields */);

N-dimensional shape of a tensor.

Implementations

impl Shape

pub fn new(dims: Vec<usize>) -> Self

Create a new shape from a vector of dimension sizes.

pub fn dims(&self) -> &[usize]

The dimension sizes as a slice.

pub fn rank(&self) -> usize

Number of dimensions (0 for scalar, 1 for vector, 2 for matrix, etc.).

pub fn elem_count(&self) -> usize

Total number of elements (product of all dimensions). A scalar shape [] has 1 element.

pub fn stride_contiguous(&self) -> Vec<usize>

Compute the contiguous (row-major / C-order) strides for this shape.

For shape [2, 3, 4], strides are [12, 4, 1]:

• Moving 1 step in dim 0 jumps 12 elements (3*4)
• Moving 1 step in dim 1 jumps 4 elements
• Moving 1 step in dim 2 jumps 1 element

This is how row-major memory works: the last dimension is contiguous.

pub fn dim(&self, d: usize) -> Result<usize>

Size of a specific dimension.

pub fn broadcast_shape(lhs: &Shape, rhs: &Shape) -> Result<Shape>

Compute the broadcast output shape from two input shapes.

NumPy-style broadcasting rules:

1. Align shapes from the right (trailing dimensions).
2. Dimensions are compatible if they are equal or one of them is 1.
3. Missing leading dimensions are treated as 1.

Examples: [3, 4] and [4] → [3, 4] (expand [4] to [1, 4] then broadcast dim 0) [2, 1] and [1, 3] → [2, 3] [5, 3, 1] and [3, 4] → [5, 3, 4] [3] and [4] → Error (3 ≠ 4 and neither is 1)

pub fn broadcast_strides(&self, target: &Shape) -> Vec<usize>

Return the broadcast strides for this shape to match a target broadcast shape.

For each dimension where self.dim[i] == 1 and target.dim[i] > 1, the stride is set to 0 (repeating the single element). For missing leading dimensions (self has fewer dims), stride is also 0.

Trait Implementations

impl Clone for Shape

fn clone(&self) -> Shape

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Shape

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for Shape

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl From<&[usize]> for Shape

fn from(s: &[usize]) -> Self

Converts to this type from the input type.

impl From<()> for Shape

fn from(_: ()) -> Self

Scalar shape (0 dimensions).

impl From<(usize,)> for Shape

fn from((d0): (usize,)) -> Self

Converts to this type from the input type.

impl From<(usize, usize)> for Shape

fn from((d0, d1): (usize, usize)) -> Self

Converts to this type from the input type.

impl From<(usize, usize, usize)> for Shape

fn from((d0, d1, d2): (usize, usize, usize)) -> Self

Converts to this type from the input type.

impl From<(usize, usize, usize, usize)> for Shape

fn from((d0, d1, d2, d3): (usize, usize, usize, usize)) -> Self

Converts to this type from the input type.

impl From<Shape> for SymbolicShape

fn from(shape: Shape) -> Self

Converts to this type from the input type.

impl From<Vec<usize>> for Shape

fn from(v: Vec<usize>) -> Self

Converts to this type from the input type.

impl From<usize> for Shape

fn from(d: usize) -> Self

1-D shape.

impl Hash for Shape

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for Shape

fn eq(&self, other: &Shape) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Eq for Shape

impl StructuralPartialEq for Shape