Big TF test cleanup (#24282)

* Fix one BLIP arg not being optional, remove misspelled arg

* Remove the lxmert test overrides and just use the base test_saved_model_creation

* saved_model_creation fixes and re-enabling tests across the board

* Remove unnecessary skip

* Stop caching sinusoidal embeddings in speech_to_text

* Fix transfo_xl compilation

* Fix transfo_xl compilation

* Fix the conditionals in xglm

* Set the save spec only when building

* Clarify comment

* Move comment correctly

* Correct embeddings generation for speech2text

* Mark RAG generation tests as @slow

* Remove redundant else:

* Add comment to clarify the save_spec line in build()

* Fix size tests for XGLM at last!

* make fixup

* Remove one band_part operation

* Mark test_keras_fit as @slow

tests/models/opt/test_modeling_tf_opt.py

@@ -20,7 +20,7 @@ import unittest
 import numpy as np
 
 from transformers import OPTConfig, is_tf_available
-from transformers.testing_utils import require_sentencepiece, require_tf, slow, tooslow
+from transformers.testing_utils import require_sentencepiece, require_tf, slow
 
 from ...test_configuration_common import ConfigTester
 from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
@@ -219,10 +219,6 @@ class TFOPTModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCase)
                             models_equal = False
                     self.assertTrue(models_equal)
 
-    @tooslow
-    def test_saved_model_creation(self):
-        pass
-
 
 def _long_tensor(tok_lst):
     return tf.constant(tok_lst, dtype=tf.int32)