/
NNFloatData

NNFloatData

Owned by 杨彬
Jan 08, 2025
2 min read
class NNFloatData : public NNData { public: static std::shared_ptr<NNFloatData> createData(uint32_t dataSize); static std::shared_ptr<NNFloatData> createData(void* data, uint32_t dataSize); static std::shared_ptr<NNFloatData> createData(std::shared_ptr<NNRawData> rawData); public: void* getData(); uint32_t getDataSize(); public: virtual std::string getType(); virtual std::string toString(); virtual std::shared_ptr<NNData> clone(bool deep); };

Introduction to the class

A class representing floating-point data within a neural network.

This class inherits from NNData and provides specific implementations for handling floating-point data. It includes methods for creating data instances, accessing the data, and fulfilling the virtual functions defined in the base class.

Introduction to the Class's Methods

  • static std::shared_ptr<NNFloatData> createData(uint32_t dataSize)

    • brief: Creates a new instance of NNFloatData with a specified data size. This method allocates memory for the floating-point data based on the provided size.

    • param

      • dataSize: The size of the data to allocate, in number of floating-point values.

    • return shared_ptr<NNFloatData>: A shared pointer to the created data instance.

  • static std::shared_ptr<NNFloatData> createData(void* data, uint32_t dataSize)

    • brief: Creates a new instance of NNFloatData from existing data. This method takes a pointer to existing data and its size, wraps it in an NNFloatData instance.

    • param

      • data: A pointer to the existing data. It should be noted that the content of the data parameter will be deep-copied into its internal properties.

      • dataSize: The size of the data, in number of floating-point values.

    • return shared_ptr<NNFloatData>: A shared pointer to the created data instance.

  • static std::shared_ptr<NNFloatData> createData(std::shared_ptr<NNRawData> rawData)

    • brief: Creates a new instance of NNFloatData from a shared pointer to raw data. This method takes a shared pointer to NNRawData and wraps it in an NNFloatData instance.

    • param

      • rawData: A shared pointer to the raw data. It should be noted that the passed-in parameter rawData will be referenced by this object.

    • return shared_ptr<NNFloatData>: A shared pointer to the created data instance.

  • void* getData()

    • brief: Gets a pointer to the internal data. This method returns a pointer to the internal floating-point data.

    • return void*: A pointer to the internal data.

  • uint32_t getDataSize()

    • brief: Gets the size of the internal data. This method returns the size of the internal floating-point data, in number of values.

    • return uint32_t: The size of the data.

  1. std::string getType()

    • brief: Gets the type of the data. This method returns a string representing the type of the data.

    • return string: A string representing the data type.

  2. std::string toString()

    • brief: Converts the data object to a string representation. This method returns a string that describes the data object, including its type and values.

    • return string: A string representation of the data object.

  3. std::shared_ptr<NNData> clone(bool deep)

    • brief: Clones the data object. This method creates a deep or shallow copy of the data object, depending on the value of the `deep` parameter.

    • param

      • deep: A boolean indicating whether to perform a deep clone.

    • return shared_ptr<NNData>: A shared pointer to the cloned data object.

Related content