A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 12, 16:00 UTC
[Misc] Fix pre-commit (#29238)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Nov 22, 2025 at 22:48 UTC
ae66818379fc2403e43c47154a98170aa7cab192
1 file changed
+1
-1
vllm/model_executor/model_loader/utils.py
+1
-1
@@ -19,7 +19,6 @@ from vllm.model_executor.layers.quantization.base_config import (
19
QuantizationConfig,
20
QuantizeMethodBase,
21
)
22
-
22
from vllm.model_executor.models.interfaces import SupportsQuant, supports_multimodal
23
from vllm.utils.platform_utils import is_pin_memory_available
24
@@ -173,6 +172,7 @@ def _get_model_architecture(model_config: ModelConfig) -> tuple[type[nn.Module],
172
as_seq_cls_model,
173
try_create_mm_pooling_model_cls,
174
)
175
+
176
architectures = getattr(model_config.hf_config, "architectures", [])
177
178
model_cls, arch = model_config.registry.resolve_model_cls(