Struct ZeroCopyTensor 

pub struct ZeroCopyTensor<T> { /* private fields */ }

Zero-copy tensor wrapper that avoids unnecessary memory allocation

This struct provides a view into existing memory without copying data. It can work with borrowed slices or take ownership of allocated memory.

Implementations

impl<T> ZeroCopyTensor<T>

pub unsafe fn from_raw_parts( data_ptr: *const T, shape: Vec<usize>, stride: Vec<usize>, ) -> Self

Create a zero-copy tensor from existing data without copying

Safety

This function is unsafe because it directly uses raw pointers. The caller must ensure:

• data_ptr is a valid pointer to a memory region that contains at least capacity elements
• The memory region remains valid for the lifetime of the ZeroCopyTensor
• The memory is properly aligned for type T
• The shape and stride parameters correctly describe the tensor layout
• No other code mutates the memory region while this tensor exists

pub fn from_slice(data: &[T], shape: Vec<usize>) -> Self

Create a zero-copy tensor from a slice

This creates a view into the provided slice without copying data. The slice must remain valid for the lifetime of the tensor.

pub fn from_vec(data: Vec<T>, shape: Vec<usize>) -> Self

Create a zero-copy tensor by taking ownership of a Vec

This transfers ownership of the Vec’s memory to the tensor, avoiding the need to copy data.

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

Get the shape of the tensor

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

Get the stride of the tensor

pub fn len(&self) -> usize

Get the total number of elements

pub fn is_empty(&self) -> bool

Check if the tensor is empty

pub fn as_slice(&self) -> &[T]

Get data as a slice

Safety

This is safe as long as the tensor was constructed properly and the underlying memory remains valid.

pub fn slice_view(&self, ranges: &[(usize, usize)]) -> Result<ZeroCopyTensor<T>>

Create a view into a subregion without copying data

This creates a new tensor that views a slice of the current tensor. The data is not copied, only the view parameters are adjusted.

pub fn ndim(&self) -> usize

Get the number of dimensions

pub fn is_owned(&self) -> bool

Check if this tensor owns its memory

Trait Implementations

impl<T> Drop for ZeroCopyTensor<T>

fn drop(&mut self)

Executes the destructor for this type.

fn pin_drop(self: Pin<&mut Self>)

🔬This is a nightly-only experimental API. (pin_ergonomics)

Execute the destructor for this type, but different to Drop::drop, it requires self to be pinned.

impl<T: Send> Send for ZeroCopyTensor<T>

impl<T: Sync> Sync for ZeroCopyTensor<T>