fix typo in GradScaler docstring (pytorch#42236)

Summary:
Closes pytorch#42226.

Pull Request resolved: pytorch#42236

Reviewed By: albanD

Differential Revision: D22817980

Pulled By: ngimel

fbshipit-source-id: 4326fe028dba1dbeed454edc4e4d4fffa56f51d6

torch/cuda/amp/grad_scaler.py

@@ -93,7 +93,7 @@ class GradScaler(object):
     Arguments:
         init_scale (float, optional, default=2.**16):  Initial scale factor.
         growth_factor (float, optional, default=2.0):  Factor by which the scale is multiplied during
-            :meth:`update` if no inf/NaN gradients occur for ``growth_factor`` consecutive iterations.
+            :meth:`update` if no inf/NaN gradients occur for ``growth_interval`` consecutive iterations.
         backoff_factor (float, optional, default=0.5):  Factor by which the scale is multiplied during
             :meth:`update` if inf/NaN gradients occur in an iteration.
         growth_interval (int, optional, default=2000):  Number of consecutive iterations without inf/NaN gradients