Class TensorData

• Defined in File TensorData.hpp

Inheritance Relationships

Derived Type

• public nvcv::TensorDataStrided (Class TensorDataStrided)

Class Documentation

class TensorData

Represents data for a tensor in the system.

The TensorData class provides an interface to access and manage the underlying data of a tensor. It offers functionalities to retrieve tensor shape, layout, data type, and other properties. The tensor’s data is encapsulated in an NVCVTensorData object.

Subclassed by nvcv::TensorDataStrided

Public Functions

inline TensorData(const NVCVTensorData &data)

Constructs a TensorData object from an NVCVTensorData instance.

Parameters:

data – The underlying tensor data representation.

inline int rank() const

Retrieves the rank (number of dimensions) of the tensor.

inline const TensorShape &shape() const &

Retrieves the shape of the tensor.

inline const TensorShape::DimType &shape(int d) const &

Retrieves a specific dimension size from the tensor shape.

Parameters:

d – The index of the dimension.

Returns:

The size of the specified dimension.

inline const TensorLayout &layout() const &

Retrieves the layout of the tensor.

inline TensorShape shape() &&

Retrieves the shape of the tensor (rvalue overload).

inline TensorShape::DimType shape(int d) &&

Retrieves a specific dimension size from the tensor shape (rvalue overload).

inline TensorLayout layout() &&

Retrieves the layout of the tensor (rvalue overload).

inline DataType dtype() const

Retrieves the data type of the tensor elements.

inline const NVCVTensorData &cdata() const &

Retrieves a constant reference to the underlying tensor data.

inline NVCVTensorData cdata() &&

Retrieves the underlying tensor data (rvalue overload).

template<typename DerivedTensorData>
Optional<DerivedTensorData> cast() const

Attempts to cast the current tensor data to a derived tensor data type.

Template Parameters:

DerivedTensorData – The derived tensor data type to cast to.

Returns:

An optional containing the casted tensor data if successful; otherwise, an empty optional.

template<typename Derived>
bool IsCompatible() const

Checks if the current tensor data is compatible with a derived type.

Template Parameters:

Derived – The derived type to check compatibility against.

Returns:

true if the current tensor data is compatible with the derived type, false otherwise.

template<typename Derived>
inline Optional<Derived> cast() const

Public Static Functions

static inline bool IsCompatibleKind(NVCVTensorBufferType kind)

Determines if a given tensor buffer type is compatible.

Parameters:

kind – The tensor buffer type to check.

Returns:

true if the buffer type is compatible, false otherwise.

Protected Functions

TensorData() = default

inline NVCVTensorData &data() &