TensorRT C++ 批量推理笔记

batch为1时，如下：

void* buffers[2];
buffers[inputIndex] = inputbuffer;
buffers[outputIndex] = outputBuffer;

但是batch_size大于1时，inputbuffer和outputBuffer如何设置，以及如何构建引擎？

批量推理准备数据：

  

   

    

     

    
    

     

      void parseYolov5(cv::Mat& img,ICudaEngine* engine,IExecutionContext* context,std::vector<Yolo::Detection>& batch_res)
     
    
   

    

     

    
    

     

      {
     
    
   

    

     

    
    

     
    // 准备数据 ---------------------------
     
    
   

    

     

    
    

     
    static float data[BATCH_SIZE * 3 * INPUT_H * INPUT_W];  //输入
     
    
   

    

     

    
    

     
    static float prob[BATCH_SIZE * OUTPUT_SIZE];            //输出
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     
    assert(engine->getNbBindings() == 2);
     
    
   

    

     

    
    

     
    void* buffers[2];
     
    
   

    

     

    
    

     
    // In order to bind the buffers, we need to know the names of the input and output tensors.
     
    
   

    

     

    
    

     
    // Note that indices are guaranteed to be less than IEngine::getNbBindings()
     
    
   

    

     

    
    

     
    const int inputIndex = engine->getBindingIndex(INPUT_BLOB_NAME);
     
    
   

    

     

    
    

     
    const int outputIndex = engine->getBindin
     
    
  
 

文章来源: blog.csdn.net，作者：AI视觉网奇，版权归原作者所有，如需转载，请联系作者。

原文链接：blog.csdn.net/jacke121/article/details/123750675