Struct safetensors::tensor::TensorView

pub struct TensorView<'data> { /* private fields */ }

A view of a Tensor within the file. Contains references to data within the full byte-buffer And is thus a readable view of a single tensor

Implementations

impl<'data> TensorView<'data>

pub fn new( dtype: Dtype, shape: Vec<usize>, data: &'data [u8] ) -> Result<Self, SafeTensorError>

Create new tensor view

pub fn dtype(&self) -> Dtype

The current tensor dtype

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

The current tensor shape

pub fn data(&self) -> &'data [u8]

The current tensor byte-buffer

pub fn sliced_data( &'data self, slices: &[TensorIndexer] ) -> Result<SliceIterator<'data>, InvalidSlice>

The various pieces of the data buffer according to the asked slice

Trait Implementations

impl<'data> Debug for TensorView<'data>

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

Formats the value using the given formatter.

impl<'data, A> IndexOp<'data, (A,)> for TensorView<'data>where A: Into<TensorIndexer>,

fn slice(&'data self, index: (A,)) -> Result<SliceIterator<'data>, InvalidSlice>

Returns a slicing iterator which are the chunks of data necessary to reconstruct the desired tensor.

impl<'data, A, B> IndexOp<'data, (A, B)> for TensorView<'data>where A: Into<TensorIndexer>, B: Into<TensorIndexer>,

fn slice( &'data self, index: (A, B) ) -> Result<SliceIterator<'data>, InvalidSlice>

Returns a slicing iterator which are the chunks of data necessary to reconstruct the desired tensor.

impl<'data, A, B, C> IndexOp<'data, (A, B, C)> for TensorView<'data>where A: Into<TensorIndexer>, B: Into<TensorIndexer>, C: Into<TensorIndexer>,

fn slice( &'data self, index: (A, B, C) ) -> Result<SliceIterator<'data>, InvalidSlice>

Returns a slicing iterator which are the chunks of data necessary to reconstruct the desired tensor.

impl<'data, A> IndexOp<'data, A> for TensorView<'data>where A: Into<TensorIndexer>,

fn slice(&'data self, index: A) -> Result<SliceIterator<'data>, InvalidSlice>

Returns a slicing iterator which are the chunks of data necessary to reconstruct the desired tensor.

impl<'data> PartialEq<TensorView<'data>> for TensorView<'data>

fn eq(&self, other: &TensorView<'data>) -> bool

This method tests for self and other values to be equal, and is used by ==.

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

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

impl<'data> View for &TensorView<'data>

fn dtype(&self) -> Dtype

The Dtype of the tensor

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

The shape of the tensor

fn data(&self) -> Cow<'_, [u8]>

The data of the tensor

fn data_len(&self) -> usize

The length of the data, in bytes. This is necessary as this might be faster to get than data().len() for instance for tensors residing in GPU.

impl<'data> View for TensorView<'data>

fn dtype(&self) -> Dtype

The Dtype of the tensor

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

The shape of the tensor

fn data(&self) -> Cow<'_, [u8]>

The data of the tensor

fn data_len(&self) -> usize

The length of the data, in bytes. This is necessary as this might be faster to get than data().len() for instance for tensors residing in GPU.

impl<'data> Eq for TensorView<'data>

impl<'data> StructuralEq for TensorView<'data>

impl<'data> StructuralPartialEq for TensorView<'data>