"Hi TensorFlow!" using API C

Question

For training purposes, how to encode this Python example using the TensorFlow C API?

import tensorflow as tf hello = tf.constant("hello TensorFlow!") sess=tf.Session() print(sess.run(hello)) 

I tried this as follows:

 #include <string.h> #include <iostream.h> #include "c_api.h" int main( int argc, char ** argv ) { TF_Graph * graph = TF_NewGraph(); TF_SessionOptions * options = TF_NewSessionOptions(); TF_Status * status = TF_NewStatus(); TF_Session * session = TF_NewSession( graph, options, status ); char hello[] = "Hello TensorFlow!"; TF_Tensor * tensor = TF_AllocateTensor( TF_STRING, 0, 0, 8 + TF_StringEncodedSize( strlen( hello ) ) ); TF_OperationDescription * operationDescription = TF_NewOperation( graph, "Const", "hello" ); TF_Operation * operation; struct TF_Output * output; TF_StringEncode( hello, strlen( hello ), 8 + ( char * ) TF_TensorData( tensor ), TF_StringEncodedSize( strlen( hello ) ), status ); TF_SetAttrTensor( operationDescription, "value", tensor, status ); TF_SetAttrType( operationDescription, "dtype", TF_TensorType( tensor ) ); operation = TF_FinishOperation( operationDescription, status ); output->oper = operation; output->index = 0; TF_SessionRun( session, 0, 0, 0, 0, // Inputs output, &tensor, 1, // Outputs &operation, 1, // Operations 0, status ); printf( "%i", TF_GetCode( status ) ); TF_CloseSession( session, status ); TF_DeleteSession( session, status ); TF_DeleteStatus( status ); TF_DeleteSessionOptions( options ); return 0; } 

I am testing it on Windows using TensorFlow.dll from: http://ci.tensorflow.org/view/Nightly/job/nightly-libtensorflow-windows/lastSuccessfulBuild/artifact/lib_package/libtensorflow-cpu-windows-x86_64.zip

The above GPF code codes are in the TF_SessionRun() call. As soon as we find a solution for this, how to get a way out? If there is a conclusion for the tensor? The above code reuses it for both output and operation.

thank you very much

Answer 1

An error occurred while initializing the offset. This version works fine:

 #include <iostream.h> #include "c_api.h" int main( int argc, char ** argv ) { TF_Graph * graph = TF_NewGraph(); TF_SessionOptions * options = TF_NewSessionOptions(); TF_Status * status = TF_NewStatus(); TF_Session * session = TF_NewSession( graph, options, status ); char hello[] = "Hello TensorFlow!"; TF_Tensor * tensor = TF_AllocateTensor( TF_STRING, 0, 0, 8 + TF_StringEncodedSize( strlen( hello ) ) ); TF_Tensor * tensorOutput; TF_OperationDescription * operationDescription = TF_NewOperation( graph, "Const", "hello" ); TF_Operation * operation; struct TF_Output output; TF_StringEncode( hello, strlen( hello ), 8 + ( char * ) TF_TensorData( tensor ), TF_StringEncodedSize( strlen( hello ) ), status ); memset( TF_TensorData( tensor ), 0, 8 ); TF_SetAttrTensor( operationDescription, "value", tensor, status ); TF_SetAttrType( operationDescription, "dtype", TF_TensorType( tensor ) ); operation = TF_FinishOperation( operationDescription, status ); output.oper = operation; output.index = 0; TF_SessionRun( session, 0, 0, 0, 0, // Inputs &output, &tensorOutput, 1, // Outputs &operation, 1, // Operations 0, status ); printf( "status code: %i\n", TF_GetCode( status ) ); printf( "%s\n", ( ( char * ) TF_TensorData( tensorOutput ) ) + 9 ); TF_CloseSession( session, status ); TF_DeleteSession( session, status ); TF_DeleteStatus( status ); TF_DeleteSessionOptions( options ); return 0; } 

Do I need to remove the tensor? Not sure why we should add 9 (instead of 8) to get the start of the line.

Answer 2

TF_STRING tensors are encoded using the format described here . In your code, you took into account a space (8 bytes) to encode one offset, but did not actually initialize it. For this you need to add something like:

memset (TF_TensorData (tensor), 0, 8);

Before calling TF_SetAttrTensor , as this will set the "offset" of the string element to 0 (where you encode one string value).

To your second question: you are not actually using the same tensor pointer. Comments for TF_SessionRun suggest that TF_SessionRun allocates a new TF_Tensor object that the call owner receives. So, in the code snippet, the tensor variable is rewritten to indicate the newly selected tensor.

Hope this helps.