Crate tenferro_tensor_core 

Backend-independent rank/layout, dtype and scalar metadata.

tenferro-tensor-core owns backend-independent tensor metadata: scalar tags and promotion facts, rank/layout metadata, and layout validation. It does not own tensor storage, execution backends, backend buffers, GPU handles, provider selection, or materializing kernels. The tensor families that carry storage live in tenferro-tensor.

Examples

use tenferro_tensor_core::{DType, Rank, TensorLayout};

let layout = TensorLayout::<Rank<2>>::compact([2, 3])?;
let transposed = layout.transpose_view([1, 0])?;
assert_eq!(transposed.shape(), &[3, 2]);

assert_eq!(DType::F64.spec().width, 64);

Macros

define_scalar_tag

Define a closed scalar set’s tag type and its promotion facts.

Structs

DynRank

Dynamic tensor rank marker.

MemberSpec

Promotion-relevant facts about one set member.

Rank

Static tensor rank marker.

SliceSpec

Explicit slice descriptor.

TensorLayout

Storage-neutral tensor layout metadata.

Enums

AdAdmissionError

Why a scalar may not be differentiated at the requested order.

DType

Runtime scalar dtype tag.

ErrorKind

Coarse classification shared by crate-local error types.

MemberKind

Kind of arithmetic a set member belongs to.

ScalarDomain

Algebra that a scalar’s ordinary arithmetic belongs to.

ShapeMismatch

Structured facts describing why two tensor shapes are incompatible.

ValidationError

Structured validation failures owned by the tensor data model.

ValidationKind

Coarse classification for shared tensor validation failures.

Traits

IntoRankShape

Convert a shape container into the representation required by a tensor rank.

IntoShapeVec

Convert a common shape container into the dynamic owned shape type.

Scalar

A scalar the host tensor data model can store and move.

ScalarArithmetic

Arithmetic a scalar supports under its own rules.

TensorRank

Rank contract for tensor metadata shapes and strides.

TensorScalar

Sealed trait for scalar types supported by the core tensor data model.

Functions

ad_admission

Answer whether the shared differentiation paths may differentiate T at order.

col_major_strides

Return compact column-major strides for a shape.

promote_in_set

Promote two members of one set from its parallel tag and fact tables.

promote_specs

Combine two member facts under the ordinary numeric promotion rules.

Type Aliases

Result

Result type for tensor data-model operations.

ShapeVec

Small tensor shape vector with inline capacity for common dynamic ranks.

StrideVec

Small tensor stride vector with signed element strides.