resolved a 0-dim tensor slicing bug from _get_state_without_padding (#7659)

fixes #7650 

adding a `value.dim()>0` check to prevent slicing of 0-dim tensors

cc @sfc-gh-truwase

Signed-off-by: Naveenraj Kamalakannan <[email protected]>
Co-authored-by: Olatunji Ruwase <[email protected]>

Author: therealnaveenkamal

deepspeed/runtime/zero/stage_1_and_2.py

@@ -2297,7 +2297,7 @@ def _get_groups_without_padding(self, groups_with_padding):
     def _get_state_without_padding(self, state_with_padding, padding):
         lean_state = {}
         for key, value in state_with_padding.items():
-            if torch.is_tensor(value):
+            if torch.is_tensor(value) and value.dim() > 0:
                 lean_length = value.numel() - padding
                 lean_state[key] = value[:lean_length]
             else: