Struct Tensor 

pub struct Tensor { /* private fields */ }

A dense, C-ordered array of numbers.

Element bytes are stored in native endianness, matching what NumPy and the tenbin format write on the same machine.

Implementations

impl Tensor

pub fn new( dtype: DType, shape: Vec<usize>, data: impl Into<Bytes>, ) -> Result<Tensor>

Build a tensor from raw native-endian element bytes.

Fails when data does not hold exactly shape.product() * dtype.size() bytes.

pub fn from_f32(values: &[f32]) -> Tensor

Build a one-dimensional f32 tensor.

pub fn from_f32_shaped(values: &[f32], shape: Vec<usize>) -> Tensor

Build an f32 tensor with an explicit shape.

Panics

Panics if shape does not describe exactly values.len() elements.

pub fn from_u8_shaped(values: Vec<u8>, shape: Vec<usize>) -> Tensor

Build a u8 tensor with an explicit shape.

Panics

Panics if shape does not describe exactly values.len() elements.

pub fn dtype(&self) -> DType

The element type.

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

The shape, outermost dimension first.

pub fn ndim(&self) -> usize

The number of dimensions.

pub fn numel(&self) -> usize

The total number of elements.

pub fn data(&self) -> &Bytes

The raw native-endian element bytes.

pub fn into_data(self) -> Bytes

Consume the tensor and return its raw element bytes.

pub fn reshape(&self, shape: Vec<usize>) -> Result<Tensor>

Change the shape, keeping the element data.

Fails when the new shape describes a different number of elements.

pub fn to_f64_vec(&self) -> Vec<f64>

Copy the elements out as f64, whatever the stored type.

pub fn to_f32_vec(&self) -> Vec<f32>

Copy the elements out as f32, whatever the stored type.

pub fn get(&self, index: usize) -> f64

Read a single element by flat index, widened to f64.

Panics

Panics if index is out of bounds.

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 Eq for Tensor

impl From<Tensor> for Value

fn from(v: Tensor) -> Value

Converts to this type from the input type.

impl PartialEq for Tensor

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for Tensor