Module zspace 

Common ZSpace Utilities for Ruda

This is a new/experimental module.

The goal will be to unify:

• shape/stride construction/validation.
• stride map view transformations.
• common Shape types.
• common Shape/Size/Reshape utility traits.

The intention is to publish this as a stand-alone zspace module, with no direct tie to ruda; once it is more polished.

Modules

dtype

Tensor data type.

errors

Common Ruda Errors

indexing

Indexing Utilities

metadata

primitive

quantization

Quantization data representation.

shape

Tensor shape definition.

slice

Tensor slice utilities.

spatial

striding

Stride Utilities

Macros

s

Slice argument constructor for tensor indexing.

shape

smallvec

Creates a SmallVec containing the arguments.

strides

Structs

BlockSize

Copyable block size, specialized version of SmallVec.

FloatInfo

Numerical precision properties for a floating-point dtype.

Metadata

QParamTensor

A quantization parameter tensor descriptor.

QParams

The quantization tensor data parameters.

QuantScheme

Describes a quantization scheme/configuration.

Shape

Shape of a tensor.

Slice

A slice specification for a single tensor dimension.

SliceIter

Defines an Iterator over a Slice.

SmallVec

A Vec-like container that can store a small number of elements inline.

Strides

Enums

BoolStore

Data type used to store boolean values.

DType

ExpressionError

Common Expression Error.

FloatDType

IntDType

MetadataError

Error that can occur when attempting to modify shapes.

QuantAcc

The precision of accumulating elements.

QuantLevel

Level or granularity of quantization.

QuantMode

Strategy used to quantize values.

QuantParam

Quantization floating-point precision.

QuantPropagation

Specify if the output of an operation is quantized using the scheme of the input or returned unquantized.

QuantStore

Data type used to stored quantized values.

QuantValue

Data type used to represent quantized values.

ReshapeAction

The action to take for a reshape operation.

ReshapeAnalysis

The reshape kind.

Constants

QPARAM_ALIGN

Alignment (in bytes) for quantization parameters in serialized tensor data.

Traits

QTensorPrimitive

Quantized tensor primitive.

SliceArg

Trait for slice arguments that can be converted into an array of slices. This allows the slice method to accept both single slices (from s![..]) and arrays of slices (from s![.., ..] or [0..5, 1..3]).

SliceOps

Slice-related ops on Shape

TensorMetadata

Tensor metadata trait for tensor primitive.

Functions

broadcast_strides

Calculate the new strides given added batch dimensions.

calculate_matmul_output

Compute the output shape for matrix multiplication with broadcasting support.

contiguous_strides

Computes the strides for a contiguous tensor with the given shape.

is_contiguous

Check if the current tensor is contiguous.

pack_i8s_to_u32s

Pack signed 8-bit integer values into a sequence of unsigned 32-bit integers.

params_shape

Calculate the shape of the quantization parameters for a given tensor and level

reshape_action

Returns the proper action to take when reshaping a tensor.

reshape_analysis

Returns the analysis of a reshape operation.

split_strides

Calculate the new strides given added split dimensions.

Type Aliases

BoolDType

Boolean dtype.