[Bugfix][Frontend] Force reasoning mode for GLM-5.3 chat templates in the GLM MoE parser (#56994)

Signed-off-by: Shijin Zhang <75300765+Dovis01@users.noreply.github.com>
Co-authored-by: Chauncey <chaunceyjiang@gmail.com>
This commit is contained in:
Shijin Zhang
2026-09-29 21:26:34 +08:00
committed by GitHub
co-authored by Chauncey
parent dfc8e0f3e2
commit d95d1dcfb9
3 changed files with 111 additions and 1 deletions
+30
View File
@@ -20,6 +20,7 @@ import regex as re
from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
from vllm.logger import init_logger
from vllm.parser.engine.events import EventType
from vllm.parser.engine.parser_engine import ParserEngine
from vllm.parser.engine.parser_engine_config import (
@@ -28,6 +29,8 @@ from vllm.parser.engine.parser_engine_config import (
Transition,
)
logger = init_logger(__name__)
if TYPE_CHECKING:
from vllm.tokenizers import TokenizerLike
from vllm.tool_parsers.abstract_tool_parser import Tool
@@ -77,6 +80,23 @@ def _glm47_arg_converter(raw_args: str, partial: bool) -> str:
return json.dumps(params, ensure_ascii=False)
def _glm53_always_thinks(tokenizer: TokenizerLike) -> bool:
# GLM-5.3 keeps the GLM-4.x prompt and tool-call format but replaces the
# enable_thinking toggle with an always-on "Reasoning Effort:" header, so
# its generation prompt always opens <think> and the model output carries
# no opening tag.
template = getattr(tokenizer, "chat_template", None)
return (
isinstance(template, str)
and "[gMASK]<sop>" in template
and "Reasoning Effort:" in template
and "enable_thinking" not in template
and "<tool_call>" in template
and "<arg_key>" in template
and "<arg_value>" in template
)
@functools.cache
def glm47_moe_config(thinking: bool = True) -> ParserEngineConfig:
arg_tag_transitions = {
@@ -192,6 +212,16 @@ class Glm47MoeParser(ParserEngine):
chat_kwargs = kwargs.get("chat_template_kwargs", {}) or {}
thinking = chat_kwargs.get("thinking", None)
enable_thinking = chat_kwargs.get("enable_thinking", None)
if (thinking is False or enable_thinking is False) and _glm53_always_thinks(
tokenizer
):
logger.warning_once(
"Ignoring enable_thinking/thinking: the GLM-5.3 chat template "
"has no thinking switch, so reasoning is always on and "
"disabling extraction would leak it into the content."
)
thinking = None
enable_thinking = None
self.thinking_enabled = (
True
if thinking is None and enable_thinking is None