[Feature] add CustomOp Decorator for UnquantizedLinearMethod and...

[Feature] add CustomOp Decorator for UnquantizedLinearMethod and UnquantizedEmbeddingMethod for vllm_hcu

vllm/model_executor/layers/linear.py

@@ -39,6 +39,7 @@ from vllm.model_executor.parameter import (
 )
 from vllm.model_executor.utils import set_weight_attrs
 from vllm.platforms import current_platform
+from vllm.model_executor.custom_op import CustomOp
 
 logger = init_logger(__name__)
 
@@ -178,8 +179,8 @@ class LinearMethodBase(QuantizeMethodBase):
         Expects create_weights to have been called before on the layer."""
         raise NotImplementedError
 
-
-class UnquantizedLinearMethod(LinearMethodBase):
+@CustomOp.register("unquantized_linear_method")
+class UnquantizedLinearMethod(LinearMethodBase, CustomOp):
     """Linear method without quantization."""
 
     def create_weights(

vllm/model_executor/layers/vocab_parallel_embedding.py

@@ -24,11 +24,12 @@ from vllm.model_executor.layers.utils import dispatch_unquantized_gemm
 from vllm.model_executor.parameter import BasevLLMParameter
 from vllm.model_executor.utils import set_weight_attrs
 from vllm.platforms import current_platform
+from vllm.model_executor.custom_op import CustomOp
 
 DEFAULT_VOCAB_PADDING_SIZE = 64
 
-
-class UnquantizedEmbeddingMethod(QuantizeMethodBase):
+@CustomOp.register("unquantized_embedding_method")
+class UnquantizedEmbeddingMethod(QuantizeMethodBase, CustomOp):
     """Unquantized method for embeddings."""
 
     def create_weights(