Struct Tensor

#[repr(C)]
pub struct Tensor { /* private fields */ }

Rust binding for the C++ type at::Tensor.

Safety

Tensor will properly manage the refcount of the underling TensorImpl.

Tensor is Send: it is safe to send across thread boundaries because the underlying C++ type is atomically refcounted.

Tensor is Sync: it can be shared across threads. The underlying C++ type has interior mutability, (i.e. a const Tensor& can be used to mutate the tensor) but we are careful to expose Rust bindings that require exclusive access (ownership or mutable reference) for any C++ code that can mutate a tensor.

Implementations

impl Tensor

pub fn device(&self) -> Device

impl Tensor

pub fn scalar_type(&self) -> ScalarType

impl Tensor

pub fn is_cuda(&self) -> bool

impl Tensor

pub fn cpu(&self) -> Tensor

impl Tensor

pub fn is_sparse(&self) -> bool

impl Tensor

pub fn is_contiguous(&self, memory_format: MemoryFormat) -> bool

impl Tensor

pub fn numel(&self) -> i64

impl Tensor

pub fn nbytes(&self) -> usize

impl Tensor

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

impl Tensor

pub fn defined(&self) -> bool

impl Tensor

pub unsafe fn data_ptr(&self) -> *const c_void

This is unsafe as it directly accesses the underlying data pointer. Additionally, this should only be used when the user is sure the tensor is defined.

pub unsafe fn mut_data_ptr(&self) -> *mut c_void

This is unsafe as it directly accesses the underlying data pointer. Additionally, this should only be used when the user is sure the tensor is defined.

pub fn copy_(&mut self, src: &Tensor)

Self-modify this tensor by copying data from another tensor. The other tensor must be the same shape as this one.

pub fn sizes(&self) -> Vec<i32>

Return the size of each dimension in this tensor.

pub fn shape(&self) -> Vec<i32>

Alias of sizes.

Trait Implementations

impl CloneUnsafe for Tensor

unsafe fn clone_unsafe(&self) -> Self

This is unsafe, it creates an alias of the underlying Tensor that is not tracked by Rust. We use this to interface with C++ functions that expect an at::Tensor.

The contract for calling this function is that the clone is local and ephemeral. More precisely:

1. The clone must not be sent to another thread (local).
2. You must guarantee that clone is dropped before the originating mutable reference is dropped (ephemeral).

impl Debug for Tensor

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Tensor

fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>

where D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Display for Tensor

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

Formats the value using the given formatter.

impl Drop for Tensor

fn drop(&mut self)

Executes the destructor for this type.

impl ExternType for Tensor

type Id = (t, o, r, c, h, (), T, e, n, s, o, r)

A type-level representation of the type’s C++ namespace and type name.

type Kind = Trivial

Either cxx::kind::Opaque or cxx::kind::Trivial.

impl From<Tensor> for IValue

fn from(value: Tensor) -> Self

Converts to this type from the input type.

impl FromPyObject<'_> for Tensor

fn extract_bound(obj: &Bound<'_, PyAny>) -> PyResult<Self>

Extracts Self from the bound smart pointer obj.

impl<'py> IntoPyObject<'py> for Tensor

type Target = PyAny

The Python output type

type Output = Bound<'py, <Tensor as IntoPyObject<'py>>::Target>

The smart pointer type to use.

type Error = PyErr

The type returned in the event of a conversion error.

fn into_pyobject(self, py: Python<'py>) -> Result<Self::Output, Self::Error>

Performs the conversion.

impl PartialEq for Tensor

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for Tensor

fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>

where S: Serializer,

Serialize this value into the given Serde serializer.

impl Send for Tensor

impl Sync for Tensor

impl VectorElement for Tensor