[Frontend] Split ServingRender into renderer and entrypoint. (#44285)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
wang.yuqi
2026-06-23 11:19:09 +00:00
committed by GitHub
parent d32575a2d2
commit 31ca9504b1
26 changed files with 1155 additions and 939 deletions
+3 -4
View File
@@ -512,12 +512,11 @@ async def test_header_dp_rank_argument():
)
# Create render serving instance (required by OpenAIServingChat)
from vllm.entrypoints.serve.render.serving import OpenAIServingRender
from vllm.renderers.online_renderer import OnlineRenderer
serving_render = OpenAIServingRender(
online_renderer = OnlineRenderer(
model_config=engine.model_config,
renderer=engine.renderer,
model_registry=models.registry,
request_logger=None,
chat_template=None,
chat_template_content_format="auto",
@@ -528,7 +527,7 @@ async def test_header_dp_rank_argument():
engine_client=engine,
models=models,
response_role="assistant",
openai_serving_render=serving_render,
online_renderer=online_renderer,
chat_template=None,
chat_template_content_format="auto",
request_logger=None,