Crate rten_tensor

Expand description

rten_tensor provides multi-dimensional arrays, commonly referred to as tensors in a machine learning context.

§Storage and layout

A tensor is a combination of data storage and a layout. The data storage determines the element type and how the data is owned. A tensor can be:

Owned (like Vec<T>)
Borrowed (like &[T] or &mut [T])
Maybe-owned (like Cow[T])

The layout determines the number of dimensions (the rank) and size of each (the shape) and how indices map to offsets in the storage. The dimension count can be static (fixed at compile time) or dynamic (variable at runtime).

§Tensor types and traits

The base type for all tensors is TensorBase. This is not normally used directly but instead via a type alias which specifies the data ownership and layout:

Rank	Owned	Borrowed	Mutably borrowed	Owned or borrowed
Static	NdTensor	NdTensorView	NdTensorViewMut	CowNdTensor
Dynamic	Tensor	TensorView	TensorViewMut	CowTensor

All tensors implement the Layout trait, which provide methods to query the shape, dimension count and strides of the tensor. Tensor views provide various methods for indexing, iterating, slicing and transforming them. The AsView trait provides access to these methods for owned and mutably borrowed tensors. Conceptually it is similar to how Deref allows accesing methods for &[T] on a Vec<T>. The preferred way to import the traits is via the prelude:

use rten_tensor::prelude::*;
use rten_tensor::NdTensor;

let tensor = NdTensor::from([[1, 2], [3, 4]]);

let transposed_elems: Vec<_> = tensor.transposed().iter().copied().collect();
assert_eq!(transposed_elems, [1, 3, 2, 4]);

§Serialization

Tensors can be serialized and deserialized using serde if the serde feature is enabled. The serialized representation of a tensor includes its shape and elements in row-major (C) order. The JSON serialization of a matrix (NdTensor<f32, 2>) looks like this for example:

{
  "shape": [2, 2],
  "data": [0.5, 1.0, 1.5, 2.0]
}

Modules§

errors: Error types that are reported by various tensor operations.
prelude: This module provides a convenient way to import the most common traits from this library via a glob import.
type_num: Traits and types for compile-time arithmetic.

Macros§

ndtensorDeprecated: Construct an NdTensor from literal elements.
tensorDeprecated: Construct a Tensor from literal elements.

Structs§

AxisChunks: Iterator over slices of a tensor along an axis. See TensorView::axis_chunks.
AxisChunksMut: Iterator over mutable slices of a tensor along an axis. See TensorViewMut::axis_chunks_mut.
AxisIter: Iterator over slices of a tensor along an axis. See TensorView::axis_iter.
AxisIterMut: Iterator over mutable slices of a tensor along an axis. See TensorViewMut::axis_iter_mut.
DynIndices: Iterator over a range of N-dimensional indices, where N is not known at compile time.
DynLayout: Defines the valid indices for an N-dimensional array and how to map them to offsets in a linear buffer, where N can be varied at runtime.
GlobalAlloc: Implementation of Alloc which wraps the global allocator.
Indices: Iterator over a range of N-dimensional indices, where N may be known at compile time (see NdIndices) or only at runtime (DynIndices).
InnerIter: Iterator over views of the N innermost dimensions of a tensor with element type T and layout L.
InnerIterMut: Iterator over mutable views of the N innermost dimensions of a tensor.
Iter: Iterator over elements of a tensor, in their logical order.
IterMut: Mutable iterator over elements of a tensor.
Lanes: Iterator over 1D slices of a tensor along a target dimension of size N.
LanesMut: Mutable version of Lanes.
NdIndices: Iterator over a range of N-dimensional indices, where N is known at compile time.
NdLayout: Defines the valid indices for an N-dimensional array and how to map them to offsets in a linear buffer, where N is known at compile time.
SliceRange: A range for slicing a Tensor or NdTensor.
TensorBase: The base type for multi-dimensional arrays. This consists of storage for elements, plus a layout which maps from a multi-dimensional array index to a storage offset. This base type is not normally used directly but instead through a type alias which selects the storage type and layout.
ViewData: Storage for an immutable tensor view.
ViewMutData: Storage for a mutable tensor view.
WeaklyCheckedView: A view of a tensor which does “weak” checking when indexing via view[<index>]. This means that it does not bounds-check individual dimensions, but does bounds-check the computed offset.

Enums§

CowData: Tensor storage which may be either owned or borrowed.
OverlapPolicy: Specifies whether a tensor or view may have an overlapping layout.
SliceItem: Specifies a subset of a dimension to include when slicing a tensor or view.

Traits§

Alloc: Storage allocation trait.
AsView: Trait implemented by all variants of TensorBase, which provides a view method to get an immutable view of the tensor, plus methods which forward to such a view.
AssumeInit: Trait for converting collections of uninitialized (MaybeUninit<T>) values to collections of corresponding initializes values (T).
IntoLayout: Trait for shapes which can be used to create a contiguous layout.
IntoSliceItems: Used to convert sequences of indices and/or ranges into a uniform [SliceItem] array that can be used to slice a tensor.
IntoStorage: Trait for converting owned and borrowed element containers (Vec<T>, slices) into their corresponding Storage type.
Layout: Layouts describe the shape of a tensor, ie. the number of dimensions and size of each, and the mapping between indices and offsets in the data storage.
MatrixLayout: Provides convenience methods for querying the shape and strides of a matrix.
MutLayout: MutLayout extends Layout with methods for creating, modifying and transforming layouts.
RandomSource: Trait for sources of random data for tensors, for use with Tensor::rand.
ResizeLayout: Trait which extends MutLayout with support for changing the number of dimensions in-place.
Scalar
Storage: Trait for backing storage used by tensors and views.
StorageMut: Trait for backing storage used by mutable tensors and views.

Functions§

is_valid_permutation: Return true if permutation is a valid permutation of dimensions for a tensor of rank ndim.
to_slice_items: Convert a slice of indices into SliceItems.

Type Aliases§

CowNdTensor: Owned or borrowed tensor with N dimensions.
CowTensor: Owned or borrowed tensor with a dynamic dimension count.
DynSliceItems: Dynamically sized array of SliceItems, which avoids allocating in the common case where the length is small.
Matrix: View of a 2D tensor.
MatrixMut: Mutable view of a 2D tensor.
NdTensor: Owned tensor with N dimensions.
NdTensorView: View of a tensor with N dimensions.
NdTensorViewMut: Mutable view of a tensor with N dimensions.
Tensor: Owned tensor with a dynamic dimension count.
TensorView: View of a tensor with a dynamic dimension count.
TensorViewMut: Mutable view of a tensor with a dynamic dimension count.

Crate rten_tensorCopy item path