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 7, 06:30 UTC
[Misc] Support OOT linear method registering (#35981)
Signed-off-by: shen-shanshan <467638484@qq.com>
Shanshan Shen committed
Mar 5, 2026 at 06:25 UTC
562339abc321ac5e86cc7b000ef0734839eea49f
1 file changed
+6
vllm/model_executor/layers/linear.py
+6
@@ -64,6 +64,12 @@ WEIGHT_LOADER_V2_SUPPORTED = [
64
]
65
66
67
+def register_weight_loader_v2_supported_method(cls):
68
+ """Decorator to register a LinearMethod as supporting weight_loader_v2."""
69
+ WEIGHT_LOADER_V2_SUPPORTED.append(cls.__name__)
70
+ return cls
71
+
72
+
73
def adjust_marlin_shard(
74
param: Parameter,
75
shard_size: int,