Report issue to ML Flow

[stevehadd]

Ml Flow correctly uses a pathlib Path object when saving a model using mlflow.keras.save_model, but it does not correctly interpret the pathlib object when loading the saved model back in using mlflow.keras.load_model.

A bug should be reported through the ML Flow issues interface. It should be very easy to create a basic code snippet to demonstrate this problem.