[Frontend][last/5] Make pooling entrypoints request schema consensus. (#31127)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
wang.yuqi
2026-02-09 06:42:38 +00:00
committed by GitHub
parent 7c233dbb36
commit 22b64948f6
24 changed files with 658 additions and 612 deletions
+1 -15
View File
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import time
from typing import Any, TypeAlias
from typing import TypeAlias
from pydantic import BaseModel, Field
@@ -23,13 +23,6 @@ from vllm.utils import random_uuid
class ScoreRequestMixin(PoolingBasicRequestMixin, ClassifyRequestMixin):
# --8<-- [start:score-extra-params]
mm_processor_kwargs: dict[str, Any] | None = Field(
default=None,
description=("Additional kwargs to pass to the HF processor."),
)
# --8<-- [end:score-extra-params]
def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams:
encoder_config = model_config.encoder_config or {}
@@ -106,13 +99,6 @@ class RerankRequest(PoolingBasicRequestMixin, ClassifyRequestMixin):
documents: ScoreInputs
top_n: int = Field(default_factory=lambda: 0)
# --8<-- [start:rerank-extra-params]
mm_processor_kwargs: dict[str, Any] | None = Field(
default=None,
description=("Additional kwargs to pass to the HF processor."),
)
# --8<-- [end:rerank-extra-params]
def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams:
encoder_config = model_config.encoder_config or {}