Skip to main content

4.2.3 Model Information Retrieval API

hbDNNGetModelNameList()​

[Function Prototype]

int32_t hbDNNGetModelNameList(const char ***modelNameList, int32_t *modelNameCount, hbPackedDNNHandle_t packedDNNHandle)

[Description]

Retrieve the list of names and the count of models pointed to by packedDNNHandle.

[Parameters]

  • [out] modelNameList : List of model names.
  • [out] modelNameCount : Number of model names.
  • [in] packedDNNHandle : D-Robotics DNN handle that points to multiple models.

[Return Type]

  • Returns 0 for a successful execution of the API, otherwise, it fails.

hbDNNGetModelHandle()​

[Function Prototype]

int32_t hbDNNGetModelHandle(hbDNNHandle_t *dnnHandle, hbPackedDNNHandle_t packedDNNHandle, const char *modelName)

[Description]

Get a handle to a model from the list of models pointed to by packedDNNHandle. The returned dnnHandle can be used across functions and threads by the caller.

[Parameters]

  • [out] dnnHandle : DNN handle that points to a model.
  • [in] packedDNNHandle : DNN handle that points to multiple models.
  • [in] modelName : Model name.

[Return Type]

  • Returns 0 for a successful execution of the API, otherwise, it fails.

hbDNNGetInputCount()[Function Prototype]​

int32_t hbDNNGetInputCount(int32_t *inputCount, hbDNNHandle_t dnnHandle)

[Description]

Get the number of input tensors for the model pointed by dnnHandle.

[Parameters]

  • [out] inputCount: The number of input tensors for the model.
  • [in] dnnHandle: DNN handle pointing to a model.

[Return Type]

  • Returns 0 if the API is executed successfully, otherwise it fails.

hbDNNGetInputName()​

[Function Prototype]

int32_t hbDNNGetInputName(const char **name, hbDNNHandle_t dnnHandle, int32_t inputIndex)

[Description]

Get the name of the input tensor for the model pointed by dnnHandle at the specified index.

[Parameters]

  • [out] name: The name of the input tensor.
  • [in] dnnHandle: DNN handle pointing to a model.
  • [in] inputIndex: The index of the input tensor.

[Return Type]

  • Returns 0 if the API is executed successfully, otherwise it fails.

hbDNNGetInputTensorProperties()​

[Function Prototype]

int32_t hbDNNGetInputTensorProperties(hbDNNTensorProperties *properties, hbDNNHandle_t dnnHandle, int32_t inputIndex)

[Description]

Get the properties of the specific input tensor for the model pointed by dnnHandle.

[Parameters]

  • [out] properties: The properties of the input tensor.
  • [in] dnnHandle: DNN handle pointing to a model.
  • [in] inputIndex: The index of the input tensor.

[Return Type]

  • Return 0 indicates successful execution of the API, otherwise it fails.

hbDNNGetOutputCount()​

[FunctionPrototype]

int32_t hbDNNGetOutputCount(int32_t *outputCount, hbDNNHandle_t dnnHandle)

[Description]

Get the number of output tensors in the model pointed by dnnHandle.

[Parameters]

  • [out] outputCount Number of output tensors in the model.
  • [in] dnnHandle DNN handle pointing to a model.

[Return Type]

  • Return 0 indicates successful execution of the API, otherwise it fails.

hbDNNGetOutputName()​

[Function Prototype]

int32_t hbDNNGetOutputName(const char **name, hbDNNHandle_t dnnHandle, int32_t outputIndex)

[Description]

Get the name of the output tensor in the model pointed by dnnHandle.

[Parameters]

  • [out] name Name of the output tensor.
  • [in] dnnHandle DNN handle pointing to a model.
  • [in] outputIndex Index of the output tensor in the model.

[Return Type]

  • Return 0 indicates successful execution of the API, otherwise it fails.

hbDNNGetOutputTensorProperties()​

【Function Prototype】

int32_t hbDNNGetOutputTensorProperties(hbDNNTensorProperties *properties, hbDNNHandle_t dnnHandle, int32_t outputIndex)

【Function Description】

Get the properties of the output tensor specified by dnnHandle in the model.

【Parameters】

  • [out] properties Information of the output tensor.
  • [in] dnnHandle DNN handle pointing to a model.
  • [in] outputIndex Index of the model's output tensor.

【Return Type】

Return 0 if the API is executed successfully, otherwise it fails.