[`CodeLlamaTokenizer`] Nit, update __init__ to make sure the AddedTokens are...

[`CodeLlamaTokenizer`] Nit, update __init__ to make sure the AddedTokens are not normalized because they are special (#27359)

* make sure tokens are properly initialized for codellama slow

* add m ore pretrained models

* style

* test more tokenizers checkpoints

src/transformers/models/code_llama/tokenization_code_llama.py

@@ -149,9 +149,9 @@ class CodeLlamaTokenizer(PreTrainedTokenizer):
     ):
         requires_backends(self, "protobuf")
         self.sp_model_kwargs = {} if sp_model_kwargs is None else sp_model_kwargs
-        bos_token = AddedToken(bos_token, lstrip=False, rstrip=False) if isinstance(bos_token, str) else bos_token
-        eos_token = AddedToken(eos_token, lstrip=False, rstrip=False) if isinstance(eos_token, str) else eos_token
-        unk_token = AddedToken(unk_token, lstrip=False, rstrip=False) if isinstance(unk_token, str) else unk_token
+        bos_token = AddedToken(bos_token, normalized=False, special=True) if isinstance(bos_token, str) else bos_token
+        eos_token = AddedToken(eos_token, normalized=False, special=True) if isinstance(eos_token, str) else eos_token
+        unk_token = AddedToken(unk_token, normalized=False, special=True) if isinstance(unk_token, str) else unk_token
 
         self.use_default_system_prompt = use_default_system_prompt
         # mark tokens special to skip them

tests/models/code_llama/test_tokenization_code_llama.py

@@ -150,6 +150,8 @@ class CodeLlamaTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
         self.tokenizers_list = [
             (self.rust_tokenizer_class, "hf-internal-testing/llama-code-tokenizer", {}),
             (self.tokenizer_class, "hf-internal-testing/llama-code-tokenizer", {}),
+            (self.tokenizer_class, "codellama/CodeLlama-34b-Instruct-hf", {}),
+            (self.rust_tokenizer_class, "codellama/CodeLlama-34b-Instruct-hf", {}),
         ]
         for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
             with self.subTest(f"{tokenizer.__class__.__name__} ({pretrained_name})"):