xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 06:25:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Fix hidden_size for multimodal classification model (#24501)

Browse Source 
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
This commit is contained in:
Jee Jee Li 2025-09-10 01:37:25 +08:00 committed by GitHub
parent b9a1c4c8a2
commit 9ad0688e43
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/model_executor/models/adapters.py
View File

@ -255,7 +255,7 @@ def as_seq_cls_model(cls: _T) -> _T:
from vllm.model_executor.models.interfaces import SupportsCrossEncoding
from vllm.sequence import IntermediateTensors
from .utils import maybe_prefix
from .utils import get_model_hidden_size, maybe_prefix
class ModelForSequenceClassification(_create_pooling_model_cls(cls),
SupportsCrossEncoding):
@ -263,9 +263,10 @@ def as_seq_cls_model(cls: _T) -> _T:
def _init_pooler(self, vllm_config: "VllmConfig", prefix: str = ""):
config = vllm_config.model_config.hf_config
quant_config = vllm_config.quant_config
hidden_size = get_model_hidden_size(config)
self.score = ReplicatedLinear(
config.hidden_size,
hidden_size,
config.num_labels,
bias=False,
params_dtype=torch.float32,

7
vllm/model_executor/models/utils.py
View File

@ -761,3 +761,10 @@ def fast_topk(values: torch.Tensor, topk: int,
else:
# Use topk for efficiency with larger k values
return torch.topk(values, topk, dim=dim)
def get_model_hidden_size(hf_config: PretrainedConfig) -> int:
if hasattr(hf_config, "hidden_size"):
return hf_config.hidden_size
text_config = hf_config.get_text_config()
return text_config.hidden_size