[Bugfix] The special_tokens in tokenizer should also be controlled by...

[Bugfix] The special_tokens in tokenizer should also be controlled by do_lower_case in encoder_config. (#20750)
Signed-off-by: wang.yuqi <noooop@126.com>

tests/tokenization/test_do_lower_case.py

+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+import pytest
+
+from vllm.transformers_utils.tokenizer import get_tokenizer
+
+TOKENIZER_NAMES = ["BAAI/bge-base-en"]
+
+
+@pytest.mark.parametrize("tokenizer_name", TOKENIZER_NAMES)
+@pytest.mark.parametrize("n_tokens", [510])
+def test_special_tokens(tokenizer_name: str, n_tokens: int):
+    tokenizer = get_tokenizer(tokenizer_name, revision="main")
+
+    prompts = '[UNK]' * n_tokens
+    prompt_token_ids = tokenizer.encode(prompts)
+    assert len(prompt_token_ids) == n_tokens + 2

vllm/transformers_utils/tokenizer.py

@@ -16,6 +16,8 @@ from transformers import (AutoTokenizer, PreTrainedTokenizer,
 
 from vllm import envs
 from vllm.logger import init_logger
+from vllm.transformers_utils.config import (
+    get_sentence_transformer_tokenizer_config)
 from vllm.transformers_utils.tokenizers import MistralTokenizer
 from vllm.transformers_utils.utils import check_gguf_file
 from vllm.utils import make_async
@@ -256,6 +258,18 @@ def get_tokenizer(
             else:
                 raise e
 
+        # The special_tokens in tokenizer should also be
+        # controlled by do_lower_case in encoder_config
+        encoder_config = get_sentence_transformer_tokenizer_config(
+            tokenizer_name, revision)
+        if isinstance(encoder_config, dict) and encoder_config.get(
+                "do_lower_case", False):
+            special_tokens_map = {
+                k: v.lower()
+                for k, v in tokenizer.special_tokens_map.items()
+            }
+            tokenizer.add_special_tokens(special_tokens_map)
+
         # NOTE: We can remove this after https://github.com/THUDM/ChatGLM3/issues/1324
         if type(tokenizer).__name__ in ("ChatGLMTokenizer",
                                         "ChatGLM4Tokenizer"):