Struct ndarray_tensorflow::NdTensor

pub struct NdTensor<T, D> where
    T: TensorType,  { /* fields omitted */ }

A wrapper for Tensor that provides views.

A Tensorflow Tensor only provides a limited API. This type is a wrapper around Tensor that makes it possible to use a tensor as an ArrayView or ArrayViewMut from the ndarray crate.

Methods

impl<T, D> NdTensor<T, D> where
    T: TensorType,
    D: Dimension, 

pub fn from_tensor(tensor: Tensor<T>) -> Result<Self, ShapeError>

Construct an ArrayTensor from a Tensor.

This function will return Err when the shape is incompatible with the shape type.

pub fn zeros<I>(shape: I) -> Self where
    I: IntoDimension<Dim = D>, 

Construct a new zero-initialized wrapped Tensor with the given shape.

pub fn inner_ref(&self) -> &Tensor<T>

Get reference to the wrapped tensor.

pub fn into_inner(self) -> Tensor<T>

Convert into the wrapped tensor.

pub fn view(&self) -> ArrayView<T, D>

Get a view of the tensor.

pub fn view_mut(&mut self) -> ArrayViewMut<T, D>

Get a mutable view of the tensor.

Trait Implementations

impl<'a, T, D> Into<ArrayBase<ViewRepr<&'a T>, D>> for &'a NdTensor<T, D> where
    T: TensorType,
    D: Dimension, 

fn into(self) -> ArrayView<'a, T, D>

Performs the conversion.

impl<'a, T, D> Into<ArrayBase<ViewRepr<&'a mut T>, D>> for &'a mut NdTensor<T, D> where
    T: TensorType,
    D: Dimension, 

fn into(self) -> ArrayViewMut<'a, T, D>

Performs the conversion.