I have AI model saved as .pb model which was trained using Tensorflow1 and was deployed in cpp code using TF1 and required dependence like bazel and gcc
now the same model needs to be used in TF2, using python we are able to load the model and do the inference with session using tf.compat.v1 but to run the pb model in cpp code there is a problem in using session
the model is getting loaded and initialized and even is session is created with ok status, while running the session to do the prediction with RUN-> command session is getting failed to run.
migration of model into saved model format is not acceptable need to use same pb model in TF2. Is there any option to run the model like python using tf.compat.v1 in cpp code to run the session
Or is there any way to load the model and use it as session in cpp code
Note: cpp code was previous running on nividia Tx2 board now same as the done in Orin Nano kit