fix: Rank is set to zero by default (#594)

Signed-off-by: Padmanabha V Seshadri <seshapad@in.ibm.com>

Author: seshapad

tuning/trainercontroller/controllermetrics/per_process_state.py

@@ -74,4 +74,4 @@ def compute(self, _: TrainerState = None, **kwargs) -> Any:
         """
         if torch.distributed.is_available() and torch.distributed.is_initialized():
             return {"rank": torch.distributed.get_rank()}
-        return {"rank": None}
+        return {"rank": 0}