Fix 2 incorrect argument positioning in Tutorial (#3042)

1. "weights_only" is augument of torch.load, not for model.load_state_dict

Here is the doc:
https://pytorch.org/docs/stable/generated/torch.load.html

Co-authored-by: Svetlana Karslioglu <[email protected]>

Author: davychxn

beginner_source/saving_loading_models.py

@@ -407,7 +407,7 @@
 # .. code:: python
 #
 #    modelB = TheModelBClass(*args, **kwargs)
-#    modelB.load_state_dict(torch.load(PATH), strict=False, weights_only=True)
+#    modelB.load_state_dict(torch.load(PATH, weights_only=True), strict=False)
 #
 # Partially loading a model or loading a partial model are common
 # scenarios when transfer learning or training a new complex model.