tensor_util.h#

Tensor conversion helpers used by schema builders, including onnx_light::ParseData() and onnx_light::ToTensor().

Declares tensor conversion helpers used by ONNX schema definitions.

This header provides utility templates to decode typed values from a Tensor wrapper and to construct TensorProto constants from scalar and vector C++ values.

namespace onnx_light

Alias that makes onnx-light headers compatible with code that references ONNX_LIGHT_NAMESPACE (the macro used in the standard onnx package).

Set to ONNX_LIGHT_NAMESPACE so both names resolve to the same namespace.

Symbol-visibility attribute for the public onnx-light C++ API.

Maps the upstream compatibility macro to onnx-light’s explicit proto ABI annotation. This keeps declarations from vendored ONNX headers visible when lib_onnx_proto uses hidden visibility by default.

Namespace alias so that ONNX C++ code (and consumers such as onnxruntime) that refers to the literal onnx namespace — rather than the ONNX_NAMESPACE macro — resolves to the onnx-light namespace. The standard onnx package lives in namespace onnx; onnx-light uses onnx_light (via ONNX_LIGHT_NAMESPACE), so this alias keeps onnx-light a true drop-in. It is only introduced when the onnx-light namespace differs from onnx.

Functions

template<typename T>
std::vector<T> ParseData(const TensorProto *tensor)#

Extracts typed element data from a Tensor wrapper.

Template Parameters:

T – Element type to decode.

Parameters:

tensor – Source tensor wrapper.

Returns:

A vector containing tensor data converted to T.

int64_t RawDataElementCount(const TensorProto &tensor, size_t element_size, bool exact_fit)#

Returns the element count implied by tensor dimensions after validating raw data size.

Parameters:

  • tensor – Source tensor.

  • element_size – Size of one decoded element.

  • exact_fit – Whether raw data must contain exactly the implied number of elements.

Returns:

The number of elements implied by the tensor dimensions.

template<typename T>
std::vector<T> ParseRawData(const TensorProto &tensor, bool exact_fit = false)#

Decodes little-endian raw tensor data as values of type T.

Template Parameters:

T – Arithmetic element type to decode.

Parameters:

  • tensor – Source tensor.

  • exact_fit – Whether raw data must contain exactly the implied number of elements.

Returns:

A vector containing the decoded elements.

template<typename T>
std::vector<T> ParseData(const Tensor *tensor)#

Extracts typed element data from a Tensor wrapper.

Template Parameters:

T – Element type to decode.

Parameters:

tensor – Source tensor wrapper.

Returns:

A vector containing tensor data converted to T.

template<typename T>
TensorProto ToTensor(const T &value)#

Creates a scalar TensorProto from a C++ value.

Template Parameters:

T – Scalar value type.

Parameters:

value – Scalar value to encode.

Returns:

A TensorProto containing one element.

template<typename T>
TensorProto ToTensor(const std::vector<T> &values)#

Creates a 1D TensorProto from a vector of values.

Template Parameters:

T – Element value type.

Parameters:

values – Values to encode.

Returns:

A TensorProto containing values in order.

inline TensorProto ToTensor(const TensorProto &tensor)#

Returns a TensorProto input unchanged.

Parameters:

tensor – TensorProto to pass through.

Returns:

tensor.