I am registering the TensorFlow model into MLFlow Model Registry. I also have a script that converts the TF model to the TF-Lite model and registers the TF-Lite model to Model Registry.
Is there a way to keep the TF and TF-lite models to be of same version number? How to deal with scenarios if model versions go out of sync? What is recommended approach to deal with such scenarios? If TF-Lite conversion fails, is there a way to use same version as TF model version registered in Model Registry?