Struct Tensor 

pub struct Tensor { /* private fields */ }

Dense, row-major tensor (shape + contiguous typed data).

Shape is listed outer-to-inner (C-order), matching typical NumPy layout.

Element dtype is always derived from the payload via Tensor::dtype.

Shape and data are private so callers cannot break the shape product == data.len() invariant after construction. Use Tensor::shape / Tensor::data accessors.

PartialEq

Equality is exact element-wise (IEEE). In particular, NaN != NaN, matching Rust’s default float PartialEq.

Serde

With the serde feature, tensors encode as { "shape": ..., "data": ... }. Deserialization always calls Tensor::new, preserving the private-field invariant. JSON is debug-only and cannot faithfully represent non-finite floats; HDF5 .nir remains the NIR interchange format.

Implementations

impl Tensor

pub fn new(shape: impl Into<Vec<usize>>, data: TensorData) -> Result<Self>

Build a tensor from shape and typed data, checking length against shape.

pub const fn dtype(&self) -> DType

Element dtype of this tensor (derived from the payload).

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

Axis lengths (outer-to-inner / C-order).

pub fn data(&self) -> &TensorData

Contiguous payload.

pub fn from_f32( shape: impl Into<Vec<usize>>, data: impl Into<Vec<f32>>, ) -> Result<Self>

f32 tensor; data.len() must equal the product of shape.

pub fn from_f64( shape: impl Into<Vec<usize>>, data: impl Into<Vec<f64>>, ) -> Result<Self>

f64 tensor; data.len() must equal the product of shape.

pub fn from_i64( shape: impl Into<Vec<usize>>, data: impl Into<Vec<i64>>, ) -> Result<Self>

i64 tensor; data.len() must equal the product of shape.

pub fn from_bool( shape: impl Into<Vec<usize>>, data: impl Into<Vec<bool>>, ) -> Result<Self>

bool tensor; data.len() must equal the product of shape.

pub fn scalar_f32(value: f32) -> Self

Rank-0 (scalar) f32 tensor.

pub fn scalar_f64(value: f64) -> Self

Rank-0 (scalar) f64 tensor.

pub fn scalar_i64(value: i64) -> Self

Rank-0 (scalar) i64 tensor.

pub fn zeros_like(&self) -> Self

A tensor of zeros with the same shape and dtype as self.

Mirrors numpy.zeros_like, which upstream NIR uses to default absent optional wire fields (v_reset). DType::Bool zeroes to false.

pub fn ones_like(&self) -> Self

A tensor of ones with the same shape and dtype as self.

Mirrors numpy.ones_like, which upstream NIR uses to default an absent w_in. DType::Bool ones to true.

pub fn numel(&self) -> usize

Number of elements implied by shape (empty shape → 1).

pub fn ndim(&self) -> usize

Number of dimensions.

Trait Implementations

impl Clone for Tensor

fn clone(&self) -> Tensor

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Tensor

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

Formats the value using the given formatter.

impl PartialEq for Tensor

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for Tensor