Misfiring tf warnings (#15442)

* Fix spurious warning in TF TokenClassification models

* Fixing one last spurious warning

* Removing outdated warning altogether

src/transformers/modeling_tf_utils.py

@@ -325,16 +325,6 @@ def booleans_processing(config, **kwargs):
                 kwargs["use_cache"] if kwargs["use_cache"] is not None else getattr(config, "use_cache", None)
             )
     else:
-        if (
-            kwargs["output_attentions"] not in (None, config.output_attentions)
-            or kwargs["output_hidden_states"] not in (None, config.output_hidden_states)
-            or ("use_cache" in kwargs and kwargs["use_cache"] not in (None, config.use_cache))
-        ):
-            tf_logger.warning(
-                "The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model. "
-                "They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`)."
-            )
-
         final_booleans["output_attentions"] = config.output_attentions
         final_booleans["output_hidden_states"] = config.output_hidden_states