Consolidate rendering parameters into RenderConfig dataclass (#24543)

Signed-off-by: sfeng33 <4florafeng@gmail.com>
This commit is contained in:
Flora Feng
2025-09-10 08:44:47 +00:00
committed by GitHub
parent feaf202e93
commit 77f62613f9
8 changed files with 167 additions and 108 deletions
+9 -4
View File
@@ -28,6 +28,7 @@ from vllm.entrypoints.openai.protocol import (ErrorResponse,
# yapf: enable
from vllm.entrypoints.openai.serving_engine import OpenAIServing
from vllm.entrypoints.openai.serving_models import OpenAIServingModels
from vllm.entrypoints.renderer import RenderConfig
from vllm.entrypoints.utils import _validate_truncation_size
from vllm.logger import init_logger
from vllm.outputs import PoolingOutput, PoolingRequestOutput
@@ -149,10 +150,7 @@ class OpenAIServingPooling(OpenAIServing):
elif isinstance(request, PoolingCompletionRequest):
engine_prompts = await renderer.render_prompt(
prompt_or_prompts=request.input,
max_length=self.max_model_len,
truncate_prompt_tokens=truncate_prompt_tokens,
add_special_tokens=request.add_special_tokens,
cache_salt=getattr(request, 'cache_salt', None),
config=self._build_render_config(request),
)
else:
raise ValueError(
@@ -270,3 +268,10 @@ class OpenAIServingPooling(OpenAIServing):
data=items,
usage=usage,
)
def _build_render_config(
self, request: PoolingCompletionRequest) -> RenderConfig:
return RenderConfig(
max_length=self.max_model_len,
truncate_prompt_tokens=request.truncate_prompt_tokens,
add_special_tokens=request.add_special_tokens)