chore: typo on ner accelerate example code (#14150)

examples/pytorch/token-classification/run_ner_no_trainer.py

@@ -99,7 +99,7 @@ def parse_args():
         default=128,
         help=(
             "The maximum total input sequence length after tokenization. Sequences longer than this will be truncated,"
-            " sequences shorter will be padded if `--pad_to_max_lenght` is passed."
+            " sequences shorter will be padded if `--pad_to_max_length` is passed."
         ),
     )
     parser.add_argument(