Struct TensorView

pub struct TensorView { /* private fields */ }

A non-owning view of a tensor: a pointer to a DLTensor.

This mirrors tvm::ffi::TensorView in C++. A function argument of this type accepts both a borrowed DLTensor* (type index kTVMFFIDLTensorPtr, how C and C++ callers pass tensors) and a Tensor object (type index kTVMFFITensor), and is passed on as a DLTensor*. Prefer it over Tensor for arguments of exported functions that only read a tensor’s metadata and data.

The view does not keep anything alive. The caller must ensure that the DLTensor it points to, and the memory that the DLTensor points to (data, shape and strides), outlive every use of the view. For an argument of an exported function, that is the duration of the call. For the same reason, as in C++, a view cannot be moved into an owned Any; use Tensor instead.

Implementations

impl TensorView

pub unsafe fn from_raw(tensor: *const DLTensor) -> Self

Creates a view of a DLTensor.

Safety

tensor must be non-null and valid for every use of the view.

pub fn as_raw(&self) -> *const DLTensor

The viewed DLTensor.

pub fn data_ptr(&self) -> *mut c_void

The data pointer, as in the DLTensor; byte_offset is not applied.

pub fn device(&self) -> DLDevice

pub fn dtype(&self) -> DLDataType

pub fn ndim(&self) -> usize

pub fn shape(&self) -> &[i64]

pub fn numel(&self) -> usize

pub fn strides(&self) -> &[i64]

The strides, in elements.

Panics

If the DLTensor has null strides and at least one dimension, which DLPack forbids since v1.2, as C++ TensorView::strides checks.

pub fn is_contiguous(&self) -> bool

Whether the elements are laid out in row-major order without gaps, as C++ tvm::ffi::IsContiguous decides: a DLTensor with null strides or no elements is contiguous, and a dimension of extent 1 may have any stride.

Trait Implementations

impl AnyCompatible for TensorView

fn type_str() -> String

the type string of the type

unsafe fn copy_to_any_view(src: &Self, data: &mut TVMFFIAny)

the value to copy to TVMFFIAny

unsafe fn move_to_any(_src: Self, _data: &mut TVMFFIAny)

consume the value to move to Any

unsafe fn check_any_strict(data: &TVMFFIAny) -> bool

unsafe fn copy_from_any_view_after_check(data: &TVMFFIAny) -> Self

Copy value from TVMFFIAny after checking caller must ensure that the value is compatible with the type

unsafe fn move_from_any_after_check(_data: &mut TVMFFIAny) -> Self

the value to move from TVMFFIAny NOTE: pay very careful attention to avoid memory leak!

unsafe fn try_cast_from_any_view(data: &TVMFFIAny) -> Result<Self, ()>

try to cast the value from AnyView

unsafe fn is_undefined(field: *const Self) -> bool

Whether field, which holds a value of this type or all-zero bytes, holds all-zero bytes that are no value of this type, as a null object reference, a string of type index None or a zero device type: a field that a creator left undefined for reflection to set, as C++ UnsafeInit leaves a reference undefined (null). Unlike a C++ undefined reference, it is no value; a nullable reference is Option<T>, whose None is a defined value.

fn get_mismatch_type_info(data: &TVMFFIAny) -> String

Get the type key of a type when TryCastFromAnyView fails.

fn type_schema() -> String

The type schema of the type: the JSON object C++ TypeTraits<T>::TypeSchema writes, which the type schemas of functions and fields embed.

fn to_any(&self) -> Any

Borrow self into an owned Any, increfing object-backed values.

impl Clone for TensorView

fn clone(&self) -> TensorView

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for TensorView

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

Formats the value using the given formatter.

impl From<&Tensor> for TensorView

fn from(tensor: &Tensor) -> Self

Converts to this type from the input type.

impl<'a> TryFrom<AnyView<'a>> for TensorView

type Error = Error

The type returned in the event of a conversion error.

fn try_from(value: AnyView<'a>) -> Result<Self>

Performs the conversion.

impl Copy for TensorView