Generate ANN weights as a separate file from Matlab embedded coder and use in VXworks environment

This question is follow up of topic in thread

https://uk.mathworks.com/matlabcentral/answers/2173605-porting-autogenerated-c-code-from-deep-learning-toolbox-to-embedded-environment?s_tid=srchtitle

The code generated has a function to read the BIN file containing neural network weights .

As shown below

Howeever, we want to generate our own handwritten function which initialises the inputBufferPtr from the file path in the OBC.

Is there way to achieve this without modifying the autocode ?

static void readDnnConstants(real32_T *inputBufferPtr,
                             const char_T *unresolvedFilePath,
                             int32_T numElementsToRead)
{
  int32_T elementSizeInBytes;
  const char_T *fileOpenMode;
  char_T *resolvedFilePath;
  FILE *filePtr;
  void *dataBufferPtr;
  resolvedFilePath = getResolvedFilePath(unresolvedFilePath);
  fileOpenMode = "rb";
  filePtr = fopen(resolvedFilePath, (char_T *)fileOpenMode);
  dataBufferPtr = &inputBufferPtr[0];
  elementSizeInBytes = 4;
  fread(dataBufferPtr, elementSizeInBytes, numElementsToRead, filePtr);
  fclose(filePtr);
  free(resolvedFilePath);
}