使用tensorflow c api运行会话比使用python要慢得多

简化,我的代码如下所示：

int main (int argc,const char * argv[])
  {
    // Initializing graph 
    tensorflow::GraphDef graph_def;
    // Folder in which graph data is located
    string graph_file_name = "Model/graph.pb";
    // Loading graph 
    tensorflow::Status graph_loaded_status =  ReadBinaryProto(tensorflow::Env::Default(),graph_file_name,&graph_def);
    if (!graph_loaded_status.ok())
    {
      cout << graph_loaded_status.ToString() << endl;
      return 1;
    }
    unique_ptr<tensorflow::Session> session_sqdet(tensorflow::NewSession(tensorflow::SessionOptions()));
    tensorflow::Status session_create_status = session_sqdet->Create(graph_def);
    if (!session_create_status.ok())
    {
      cout << "Session create status: fail." << endl;
      return 1;
    }
    while ()
    {
      /* create & preprocess batch */

      session.Run({{ "image_input",input_tensor},{"keep_prob",prob_tensor}},{"probability/score","bbox/trimming/bbox"},{},&final_output);

      /* do some postprocessing */
    }
  }

我尝试过的：

1)使用优化标志 – 所有都打开,没有警告.

2)使用批处理：性能提高,但python和C之间的差距仍然很大(运行会话需要1s vs 2.4s,batch_size = 20).

任何帮助将受到高度赞赏.