Fix past CI (#20967)

* Fix for Past CI

* make style

* clean up

* unindent 2 blocks

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

tests/models/bert/test_tokenization_bert_tf.py

@@ -4,15 +4,15 @@ from tempfile import TemporaryDirectory
 
 from transformers import AutoConfig, TFAutoModel, is_tensorflow_text_available, is_tf_available
 from transformers.models.bert.tokenization_bert import BertTokenizer
-from transformers.testing_utils import require_tensorflow_text, slow
+from transformers.testing_utils import require_tensorflow_text, require_tf, slow
 
 
-if is_tensorflow_text_available():
-    from transformers.models.bert import TFBertTokenizer
-
 if is_tf_available():
     import tensorflow as tf
 
+if is_tensorflow_text_available():
+    from transformers.models.bert import TFBertTokenizer
+
 
 TOKENIZER_CHECKPOINTS = ["bert-base-uncased", "bert-base-cased"]
 TINY_MODEL_CHECKPOINT = "hf-internal-testing/tiny-bert-tf-only"
@@ -32,6 +32,7 @@ if is_tf_available():
             return out["pooler_output"]
 
 
+@require_tf
 @require_tensorflow_text
 class BertTokenizationTest(unittest.TestCase):
     # The TF tokenizers are usually going to be used as pretrained tokenizers from existing model checkpoints,

tests/models/gpt2/test_tokenization_gpt2_tf.py

@@ -4,15 +4,15 @@ from tempfile import TemporaryDirectory
 
 from transformers import AutoConfig, TFGPT2LMHeadModel, is_keras_nlp_available, is_tf_available
 from transformers.models.gpt2.tokenization_gpt2 import GPT2Tokenizer
-from transformers.testing_utils import require_keras_nlp, slow
+from transformers.testing_utils import require_keras_nlp, require_tf, slow
 
 
-if is_keras_nlp_available():
-    from transformers.models.gpt2 import TFGPT2Tokenizer
-
 if is_tf_available():
     import tensorflow as tf
 
+if is_keras_nlp_available():
+    from transformers.models.gpt2 import TFGPT2Tokenizer
+
 
 TOKENIZER_CHECKPOINTS = ["gpt2"]
 TINY_MODEL_CHECKPOINT = "gpt2"
@@ -40,6 +40,7 @@ if is_tf_available():
             return outputs
 
 
+@require_tf
 @require_keras_nlp
 class GPTTokenizationTest(unittest.TestCase):
     # The TF tokenizers are usually going to be used as pretrained tokenizers from existing model checkpoints,

tests/models/layoutlmv2/test_tokenization_layoutlmv2.py

@@ -38,7 +38,14 @@ from transformers.models.layoutlmv2.tokenization_layoutlmv2 import (
     _is_punctuation,
     _is_whitespace,
 )
-from transformers.testing_utils import is_pt_tf_cross_test, require_pandas, require_tokenizers, require_torch, slow
+from transformers.testing_utils import (
+    is_pt_tf_cross_test,
+    require_detectron2,
+    require_pandas,
+    require_tokenizers,
+    require_torch,
+    slow,
+)
 
 from ...test_tokenization_common import (
     SMALL_TRAINING_CORPUS,
@@ -1264,6 +1271,7 @@ class LayoutLMv2TokenizationTest(TokenizerTesterMixin, unittest.TestCase):
                 self.assertEqual(sum(tokens_with_offsets["special_tokens_mask"]), added_tokens)
 
     @require_torch
+    @require_detectron2
     @slow
     def test_torch_encode_plus_sent_to_model(self):
         import torch