Use lru_cache for tokenization tests (#36818)

* fix

* fix

* fix

* fix

---------

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

tests/models/tapas/test_tokenization_tapas.py

@@ -112,8 +112,9 @@ class TapasTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
 
         return output_txt, output_ids
 
-    def setUp(self):
-        super().setUp()
+    @classmethod
+    def setUpClass(cls):
+        super().setUpClass()
 
         vocab_tokens = [
             "[UNK]",
@@ -132,8 +133,8 @@ class TapasTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
             "low",
             "lowest",
         ]
-        self.vocab_file = os.path.join(self.tmpdirname, VOCAB_FILES_NAMES["vocab_file"])
-        with open(self.vocab_file, "w", encoding="utf-8") as vocab_writer:
+        cls.vocab_file = os.path.join(cls.tmpdirname, VOCAB_FILES_NAMES["vocab_file"])
+        with open(cls.vocab_file, "w", encoding="utf-8") as vocab_writer:
             vocab_writer.write("".join([x + "\n" for x in vocab_tokens]))
 
     def get_input_output_texts(self, tokenizer):
@@ -352,7 +353,7 @@ class TapasTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
     def test_offsets_with_special_characters(self):
         for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
             with self.subTest(f"{tokenizer.__class__.__name__} ({pretrained_name})"):
-                tokenizer_r = self.rust_tokenizer_class.from_pretrained(pretrained_name, **kwargs)
+                tokenizer_r = self.get_rust_tokenizer(pretrained_name, **kwargs)
 
                 sentence = f"A, naïve {tokenizer_r.mask_token} AllenNLP sentence."
                 tokens = tokenizer_r.encode_plus(