biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Chore] Cleanup guided namespace, move to structured outputs config (#22772)

Browse Source 
Signed-off-by: Aaron Pham <contact@aarnphm.xyz>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Aaron Pham
2025-09-18 05:20:27 -04:00
committed by GitHub
parent 05b044e698
commit 29283e8976
51 changed files with 579 additions and 806 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/transformers_utils/tokenizers/mistral.py
View File

@@ -274,7 +274,7 @@ class MistralTokenizer(TokenizerBase):
return tokenizer_file
# the following attributes are set to fit vLLM's design and are used
# by the guided structured output backends.
# by the structured output backends.
@property
def all_special_tokens_extended(self) -> list[str]:
from mistral_common.tokens.tokenizers.base import SpecialTokens
@@ -463,9 +463,6 @@ class MistralTokenizer(TokenizerBase):
return decoded
# WARN: Outlines logits processors can overwrite this method.
# See: guided_decoding/outlines_logits_processors.py::_adapt_tokenizer
# for more.
def decode(self,
ids: Union[list[int], int],
skip_special_tokens: bool = True) -> str: