diff --git a/fairscale/nn/misc/flatten_params_wrapper.py b/fairscale/nn/misc/flatten_params_wrapper.py
index bc1209bbb..30b88360d 100644
--- a/fairscale/nn/misc/flatten_params_wrapper.py
+++ b/fairscale/nn/misc/flatten_params_wrapper.py
@@ -372,7 +372,7 @@ def _unflatten_params_as_views(self) -> None:
         self.flat_param unchanged.
         """
         assert self.is_flattened
-        #logger.info(f"CHRISLOG: {self._require_backward_grad_sync=}")
+        # logger.info(f"CHRISLOG: {self._require_backward_grad_sync=}")
         if self._require_backward_grad_sync:
             #logger.info("CHRISLOG: calling self.get_param_views() without torch.no_grad()")
             ps = self.get_param_views()
@@ -385,7 +385,7 @@ def _unflatten_params_as_views(self) -> None:
         for (_, m, n), p in zip(self._param_infos, ps):
             setattr(p, '_fsdp_weight', True)
             setattr(m, n, p)  # This will set as plain attr
-            #logger.info(f"CHRISLOG: {n=}, {p.requires_grad=}")
+            # logger.info(f"CHRISLOG: {n=}, {p.requires_grad=}, {p.grad_fn=}")
             param_views.append(p)
 
         # Save param views for easy access if anyone still wants to access