Inference engine handle. After the user initializes the model file information,
the SDK will automatically allocate inference engine resources according to the device conditions
to complete registration and initialization. More...

#include <native_vpu_api.h>

Public Attributes
void *	parent

char	dev_sn [64]

char	pipe_name [128]

char	model_path [128]

char	json_path [128]

Detailed Description

Inference engine handle. After the user initializes the model file information,
the SDK will automatically allocate inference engine resources according to the device conditions
to complete registration and initialization.

NccPipeHandle_t handle = { NULL,\
                             "auto",\
                             "face-detection-retail-0004",\
                             "/usr/lib/openncc/model_zoo/ncc/openvino_2021.4/face-detection-retail-0004/face-detection-retail-0004.blob",\
                             "/usr/lib/openncc/model_zoo/ncc/openvino_2021.4/face-detection-retail-0004/config/input_BGR.json"};
ret=ncc_pipe_create(&handle, NCC_SYNC);
if(ret>0)
{
    printf("ncc_pipe_create error %d !\n", ret);
    return -1;
}

Member Data Documentation

◆ dev_sn

char NccPipeHandle_t::dev_sn[64]

Parent handle to the engine, read-only

◆ json_path

char NccPipeHandle_t::json_path[128]

< AI model file path to download and inference

◆ model_path

char NccPipeHandle_t::model_path[128]

< User named pipe name.

◆ parent

void* NccPipeHandle_t::parent

◆ pipe_name

char NccPipeHandle_t::pipe_name[128]

< Assigned engine's serial number