Update vllm_omni/benchmarks/patch/patch.py

Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Signed-off-by: Zhou Taichang <tzhouam@connect.ust.hk>

vllm_omni/benchmarks/patch/patch.py

@@ -313,7 +313,7 @@ async def benchmark(
     base_url: str,
     model_id: str,
     model_name: str,
-    tokenizer: PreTrainedTokenizerBase | None,
+    tokenizer: TokenizerLike | None,
     input_requests: list[SampleRequest],
     logprobs: int | None,
     request_rate: float,