I have an already compiled model from a vendor. This model contains C++ code and unserializable contents. But I still want to register the model to MLFlow for model management. I am unable to do so with save_model or log_model in MLFlow. Any way to resolve this?
I tried to save and log the model using the API but an error is thrown when it tries to serialize the already binary model: ValueError: ctypes objects containing pointers cannot be pickled