[Frontend] Complete OpenAI render delegation (#37287)

Signed-off-by: Sage Ahrac <sagiahrak@gmail.com>
This commit is contained in:
Sage
2026-03-17 13:53:55 +00:00
committed by GitHub
parent 56cb1baa66
commit 59192dfd39
8 changed files with 140 additions and 244 deletions
+1
View File
@@ -68,6 +68,7 @@ def init_pooling_state(
OpenAIServingPooling(
engine_client,
state.openai_serving_models,
state.openai_serving_render,
request_logger=request_logger,
chat_template=resolved_chat_template,
chat_template_content_format=args.chat_template_content_format,
+7 -4
View File
@@ -32,6 +32,7 @@ from vllm.entrypoints.pooling.utils import (
encode_pooling_output_base64,
encode_pooling_output_float,
)
from vllm.entrypoints.serve.render.serving import OpenAIServingRender
from vllm.inputs import ProcessorInputs
from vllm.logger import init_logger
from vllm.outputs import PoolingRequestOutput
@@ -47,6 +48,7 @@ class OpenAIServingPooling(OpenAIServing):
self,
engine_client: EngineClient,
models: OpenAIServingModels,
openai_serving_render: OpenAIServingRender,
*,
request_logger: RequestLogger | None,
chat_template: str | None,
@@ -59,6 +61,7 @@ class OpenAIServingPooling(OpenAIServing):
request_logger=request_logger,
)
self.openai_serving_render = openai_serving_render
self.chat_template = chat_template
self.chat_template_content_format: Final = chat_template_content_format
self.trust_request_chat_template = trust_request_chat_template
@@ -101,12 +104,12 @@ class OpenAIServingPooling(OpenAIServing):
raw_prompts = await self.io_processor.pre_process_async(
prompt=validated_prompt, request_id=request_id
)
engine_prompts = await self._preprocess_cmpl(
engine_prompts = await self.openai_serving_render.preprocess_cmpl(
request,
prompt_to_seq(raw_prompts),
)
elif isinstance(request, PoolingChatRequest):
error_check_ret = self._validate_chat_template(
error_check_ret = self.openai_serving_render.validate_chat_template(
request_chat_template=request.chat_template,
chat_template_kwargs=request.chat_template_kwargs,
trust_request_chat_template=self.trust_request_chat_template,
@@ -114,7 +117,7 @@ class OpenAIServingPooling(OpenAIServing):
if error_check_ret is not None:
return error_check_ret
_, engine_prompts = await self._preprocess_chat(
_, engine_prompts = await self.openai_serving_render.preprocess_chat(
request,
request.messages,
default_template=self.chat_template,
@@ -122,7 +125,7 @@ class OpenAIServingPooling(OpenAIServing):
default_template_kwargs=None,
)
elif isinstance(request, PoolingCompletionRequest):
engine_prompts = await self._preprocess_completion(
engine_prompts = await self.openai_serving_render.preprocess_completion(
request,
prompt_input=request.input,
prompt_embeds=None,