From 4c7629cae94d1a4a8ba91d16946bbc283ecd3413 Mon Sep 17 00:00:00 2001 From: Aaron Pham Date: Sat, 15 Mar 2025 01:09:51 -0400 Subject: [PATCH] [V1][Structured Output] calculate vocab_size eagerly (#14851) Signed-off-by: Aaron Pham --- vllm/v1/structured_output/__init__.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/vllm/v1/structured_output/__init__.py b/vllm/v1/structured_output/__init__.py index 32ea1852d0ac2..77bafdee85ce2 100644 --- a/vllm/v1/structured_output/__init__.py +++ b/vllm/v1/structured_output/__init__.py @@ -40,7 +40,7 @@ class StructuredOutputManager: tokenizer_group.ping() tokenizer = tokenizer_group.get_lora_tokenizer(None) - self.vocab_size = tokenizer.max_token_id + 1 + self.vocab_size = len(tokenizer.get_vocab()) if isinstance(tokenizer, MistralTokenizer): # NOTE: ideally, xgrammar should handle this accordingly. # refer to https://github.com/mlc-ai/xgrammar/blob/d77c0a0173ef14779c918e3be7966ba852f7910f/python/xgrammar/tokenizer_info.py#L98