Unable to restore custom metric when init_from_checkpoints is used

Since the default metric is dice, resuming from a previously trained checkpoint won't pose any issue. However, if the model was initially trained using a different metric or loss. Then, we have to store that information while `saving` to make loading easier.

For more info about this, please visit https://gablab.slack.com/archives/G01DLL64143/p1715631426650549