biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Model][7/N] Improve all pooling task | Deprecation as_reward_model. Extract hidden states prefer using new multi-vector retrieval API (#26686)

Browse Source 
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
wang.yuqi
2025-12-08 16:10:09 +08:00
committed by GitHub
parent bcb6f5947f
commit 9e77ffca3f
8 changed files with 88 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/models/test_registry.py
View File

@@ -13,7 +13,6 @@ from vllm.model_executor.models import (
)
from vllm.model_executor.models.adapters import (
as_embedding_model,
as_reward_model,
as_seq_cls_model,
)
from vllm.model_executor.models.registry import (
@@ -46,7 +45,6 @@ def test_registry_imports(model_arch):
# All vLLM models should be convertible to a pooling model
assert is_pooling_model(as_seq_cls_model(model_cls))
assert is_pooling_model(as_embedding_model(model_cls))
assert is_pooling_model(as_reward_model(model_cls))
if model_arch in _MULTIMODAL_MODELS:
assert supports_multimodal(model_cls)

2
tests/test_config.py
View File

@@ -97,7 +97,7 @@ def test_update_config():
("intfloat/multilingual-e5-small", "pooling", "none", "embed"),
("jason9693/Qwen2.5-1.5B-apeach", "pooling", "classify", "classify"),
("cross-encoder/ms-marco-MiniLM-L-6-v2", "pooling", "none", "classify"),
("Qwen/Qwen2.5-Math-RM-72B", "pooling", "none", "reward"),
("Qwen/Qwen2.5-Math-RM-72B", "pooling", "none", "embed"),
("openai/whisper-small", "generate", "none", "transcription"),
],
)