xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-20 04:27:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

[V1][Structured Output] calculate vocab_size eagerly (#14851)

Browse Source 
Signed-off-by: Aaron Pham <contact@aarnphm.xyz>
This commit is contained in:
Aaron Pham 2025-03-15 01:09:51 -04:00 committed by GitHub
parent e0fdfa1608
commit 4c7629cae9
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/v1/structured_output/__init__.py
View File

@ -40,7 +40,7 @@ class StructuredOutputManager:
tokenizer_group.ping() tokenizer_group.ping()
tokenizer = tokenizer_group.get_lora_tokenizer(None) 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): if isinstance(tokenizer, MistralTokenizer):
# NOTE: ideally, xgrammar should handle this accordingly. # NOTE: ideally, xgrammar should handle this accordingly.
# refer to https://github.com/mlc-ai/xgrammar/blob/d77c0a0173ef14779c918e3be7966ba852f7910f/python/xgrammar/tokenizer_info.py#L98 # refer to https://github.com/mlc-ai/xgrammar/blob/d77c0a0173ef14779c918e3be7966ba852f7910f/python/xgrammar/tokenizer_info.py#L98