Fix distributor class name typo in docstring (#190)

Summary:
Pull Request resolved: #190

As titled.

Reviewed By: hjmshi

Differential Revision: D76954000

fbshipit-source-id: 3524c032925f381d44d417907941868c148e0e09

Author: haoyuz

distributed_shampoo/utils/shampoo_distributor.py

@@ -140,9 +140,9 @@ def _get_params_or_grads(
     def _get_params_or_grads(self, get_grad: bool = False) -> Iterable[Tensor | None]:
         """Helper function that gets params or grads from the parameter group.
 
-        NOTE: The purpose of this function is for FullyShardShampooDistributor (supporting
-        Shampoo on per-parameter FSDP, a.k.a. FSDP2 or FullyShard) to override, in order to
-        get the local params/grads from DTensors.
+        NOTE: The purpose of this function is for FullyShardDistributor (supporting Shampoo on
+        per-parameter FSDP, a.k.a. FSDP2 or FullyShard) to override, in order to get the local
+        params/grads from DTensors.
 
         By default, we just return the original params/grads.