Htykuut

The python code runs fine, but the c++ code needs some fixing:
Tensorflow .pb file in c++, opencv, placeholder tensor 'y_true' with dtype float and shape [?,3]

I am having trouble getting the correct formats for c++. I can run the code in python.
One error message I receive in c++ is: You must feed a value for placeholder tensor 'y_true' with dtype float and shape [?,3]
[[Node: y_true = Placeholderdtype=DT_FLOAT, shape=[?,3], _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
The error message is because the result is not returned to the correct size array in c++, but I am having trouble understanding how to fix this.

The code is below: the lines I'm not understanding are where to put y_test_images in the c++ code:
pythyon:

feed_dict_testing = {x: x_batch, y_true: y_test_images}

result=sess.run(y_pred, feed_dict=feed_dict_testing)

c++:
I think I should put t somewhere, but not sure where:

tensorflow::Tensor t(tensorflow::DT_FLOAT, tensorflow::TensorShape({1, 3}));

This is the line that is not returning the correct result

   tensorflow::Status run_status  = session_inception->Run({{InputName,input_tensor}},{{OutputName}},{},&finalOutput);

In python the following code works:

sess = tf.Session()

saver = tf.train.import_meta_graph('trained\break_up.meta')



sess.run(tf.global_variables_initializer())

saver.restore(sess, tf.train.latest_checkpoint('trained'))

graph = tf.get_default_graph()

y_pred = graph.get_tensor_by_name("y_pred:0")



x= graph.get_tensor_by_name("x:0") 

y_true = graph.get_tensor_by_name("y_true:0") 

y_test_images = np.zeros((1, 3)) 



images2 = np.array(images_tf, dtype=np.uint8)

images2 = images2.astype('float32')

images2 = np.multiply(images2, 1.0/255.0)

x_batch = images2.reshape(1, 32,32,3)

feed_dict_testing = {x: x_batch, y_true: y_test_images}

result=sess.run(y_pred, feed_dict=feed_dict_testing)

print(result[0,0],result[0,1],result[0,2])

I am having trouble with the y_test_images numpy array on the c++ side.

  tensorflow::GraphDef graph_def;



  std::string graphFile = "break_up2.pb";



  tensorflow::Status graphLoadedStatus = ReadBinaryProto(tensorflow::Env::Default(),graphFile,&graph_def);

  if (!graphLoadedStatus.ok()){

    std::cout << graphLoadedStatus.ToString()<<std::endl;

    return 1;

  }

  int node_count = graph_def.node_size();

  for (int i = 0; i < node_count; i++)

  {

          auto n = graph_def.node(i);

          std::cout<<"Names : "<< n.name() <<std::endl;

          std::cout<<"OP : "<< n.op() <<std::endl;



  }



  std::unique_ptr<tensorflow::Session> session_inception(tensorflow::NewSession(tensorflow::SessionOptions()));

  tensorflow::Status session_create_status = session_inception->Create(graph_def);

  if (!session_create_status.ok()){

      std::cout << session_create_status.ToString()<<std::endl;

    return 1;

  }



  std::vector<tensorflow::Tensor> finalOutput;



  std::string InputName = "x:0";

  std::string OutputName = "y_true:0";

  std::vector<float> data_y[3];

  tensorflow::Tensor t(tensorflow::DT_FLOAT, tensorflow::TensorShape({1, 3}));

  auto t_matrix = t.matrix<float>();

  t_matrix(0, 0) = 1.0;

  t_matrix(0, 1) = 0.5;

  t_matrix(0, 2) = 0.0;



  tensorflow::Status run_status  = session_inception->Run({{InputName,input_tensor}},{{OutputName}},{},&finalOutput);

  if (!run_status.ok()){

      std::cout << run_status.ToString() << "n";

    return 1;

  }