Fix typo under torch/_dynamo directory (#110459)

Summary:
This PR fixes typo of comments in files under `torch/_dynamo` directory

X-link: pytorch/pytorch#110459
Approved by: https://github.com/colesbury

Reviewed By: PaliC

Differential Revision: D49919270

fbshipit-source-id: 7c688ed0a529f39ea61a545021fc924ede508909

userbenchmark/dynamo/dynamobench/_dynamo/utils.py

@@ -1424,7 +1424,7 @@ def run_node(tracer, node, args, kwargs, nnmodule):
     """
     Runs a given node, with the given args and kwargs.
 
-    Behavior is dicatated by a node's op.
+    Behavior is dictated by a node's op.
 
     run_node is useful for extracting real values out of nodes.
     See get_real_value for more info on common usage.
@@ -1593,7 +1593,7 @@ def tensor_always_has_static_shape(
 
     Args:
     tensor - the real tensor to evaluate, parameters force a static shape.
-    is_tensor - internal dynamo check, esentially "is_tensor": target_cls is TensorVariable,
+    is_tensor - internal dynamo check, essentially "is_tensor": target_cls is TensorVariable,
     tensors not in a TensorVariable for whatever reason are forced static.
 
     Returns a tuple, where the first element is the bool of whether or not this tensor should have a static shape.
@@ -1832,7 +1832,7 @@ def defake(x):
 
 
 def is_utils_checkpoint(obj):
-    # Lazy import to avoid circular dependenices
+    # Lazy import to avoid circular dependencies
     import torch.utils.checkpoint
 
     return obj is torch.utils.checkpoint.checkpoint
@@ -1842,8 +1842,8 @@ def build_checkpoint_variable(**options):
     import torch._higher_order_ops.wrap as higher_order_ops
     from .variables.higher_order_ops import TorchHigherOrderOperatorVariable
 
-    # TODO - This is a temporary sitaution where we have two versions of
-    # checkpointing implemetation. We will converge on one and remove the other.
+    # TODO - This is a temporary situation where we have two versions of
+    # checkpointing implementation. We will converge on one and remove the other.
     activation_checkpoint_op = higher_order_ops.tag_activation_checkpoint
     if torch._functorch.config.functionalize_rng_ops:
         activation_checkpoint_op = higher_order_ops.wrap_activation_checkpoint