protocol buffers - Get tensorflow serving response directly in output proto format

Question

I have an output vector coming from tensorflow model which is a 50x1 vector of probabilities [0.4 0.3 0.4 ... 0.5] where first 20 dimensions are user vectors and last 30 are item vectors each with a particular proto value. Is there a way to directly export proto values from within tensorflow as part of tfserving For example

enum User {
    Feature1 = 0;
    // Geo
    Feature2 = 1101;
    Feature3 = 1102;
   
    Feature20}

enum Item{
   Feature1 = 0
    

   Feature30 = 2103
}
message UserProb {
User user = 1;
float confidence = 2;

}

      message ItemProb {
    Item user = 1;
    float confidence = 2;
}

question from:https://stackoverflow.com/questions/65905988/get-tensorflow-serving-response-directly-in-output-proto-format

Answer 1

Waitting for answers