Accept closure argument in NGD optimizer step (#2118)

* Accept closure argument in NGD optimizer `step`

Make this consistent with the base optimizer class

* Update NGD docs.

Co-authored-by: Geoff Pleiss <[email protected]>

Author: dannyfriar

gpytorch/optim/ngd.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 
-from typing import Iterable, Union
+from typing import Callable, Iterable, Optional, Union
 
 import torch
 
@@ -28,8 +28,13 @@ def __init__(self, params: Iterable[Union[torch.nn.Parameter, dict]], num_data:
         super().__init__(params, defaults=dict(lr=lr))
 
     @torch.no_grad()
-    def step(self) -> None:
-        """Performs a single optimization step."""
+    def step(self, closure: Optional[Callable] = None) -> None:
+        """
+        Performs a single optimization step.
+
+        (Note that the :attr:`closure` argument is not used by this optimizer; it is simply included to be
+        compatible with the PyTorch optimizer API.)
+        """
         for group in self.param_groups:
             for p in group["params"]:
                 if p.grad is None: