apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 08:21:53 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[FDConfig]Remove reasoning_parser/guided_decoding_backend/disable_any_whitespace/device_ids in FDConfig (#4362)

Browse Source 
* remove devices id

* fix unittest

* fix ce

---------

Co-authored-by: root <root@yqlcc01-sys-rpm12rzmwjd.yqlcc01.baidu.com>
This commit is contained in:
YuanRisheng
2025-10-17 10:40:59 +08:00
committed by GitHub
parent d1637db86a
commit a37c9416ac
14 changed files with 73 additions and 59 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/guided_decoding/__init__.py
+2 -2
View File
@@ -41,7 +41,7 @@ def get_guided_backend(
Raises:
ValueError: If the specified backend is not supported
"""
if fd_config.parallel_config.guided_decoding_backend.lower() == "xgrammar":
if fd_config.structured_outputs_config.guided_decoding_backend.lower() == "xgrammar":
from fastdeploy.model_executor.guided_decoding.xgrammar_backend import (
XGrammarBackend,
)
@@ -52,7 +52,7 @@ def get_guided_backend(
)
else:
raise ValueError(
f"Get unsupported backend {fd_config.parallel_config.guided_decoding_backend},"
f"Get unsupported backend {fd_config.structured_outputs_config.guided_decoding_backend},"
f" please check your configuration."
)
fastdeploy/model_executor/guided_decoding/base_guided_decoding.py
+2 -2
View File
@@ -142,9 +142,9 @@ class BackendBase:
self.reasoning_parser = None
self.hf_tokenizer = self._get_tokenizer_hf()
if self.fd_config.model_config.reasoning_parser:
if self.fd_config.structured_outputs_config.reasoning_parser:
reasoning_parser_obj = ReasoningParserManager.get_reasoning_parser(
self.fd_config.model_config.reasoning_parser
self.fd_config.structured_outputs_config.reasoning_parser
)
self.reasoning_parser = reasoning_parser_obj(self.hf_tokenizer)
fastdeploy/model_executor/guided_decoding/xgrammar_backend.py
+1 -1
View File
@@ -212,7 +212,7 @@ class XGrammarBackend(BackendBase):
self.vocab_size = fd_config.model_config.vocab_size
self.batch_size = fd_config.scheduler_config.max_num_seqs
self.any_whitespace = not fd_config.parallel_config.disable_any_whitespace
self.any_whitespace = not fd_config.structured_outputs_config.disable_any_whitespace
try:
tokenizer_info = TokenizerInfo.from_huggingface(self.hf_tokenizer, vocab_size=self.vocab_size)