[BugFix] fix qwen3-embedding model tp>1

fastdeploy/model_executor/layers/embeddings.py

@@ -164,7 +164,9 @@ def __init__(
                 ),
             )
             if self.world_size > 1:
-                set_weight_attrs(self.embeddings.weight, {"output_dim": False, "weight_loader": self.weight_loader})
+                set_weight_attrs(self.embeddings.weight, {"output_dim": False})
+                if num_embeddings % self.world_size != 0:
+                    set_weight_attrs(self.embeddings.weight, {"weight_loader", self.weight_loader})
         else:
             # column cut embedding
             self.embeddings = nn.Embedding(
@@ -236,6 +238,9 @@ def weight_loader(self, param, loaded_weight, shard_id=None):
         output_dim = getattr(param, "output_dim", None)
         packed_dim = getattr(param, "packed_dim", None)
 
+        if not param._is_initialized():
+            param.initialize()
+
         loaded_weight = get_tensor(loaded_weight)
         if param.dtype != loaded_weight.dtype:
             if loaded_weight.dtype == paddle.int8 and param.dtype == paddle.float8_e4m3fn:
@@ -247,7 +252,7 @@ def weight_loader(self, param, loaded_weight, shard_id=None):
             assert (
                 param.shape == loaded_weight.shape
             ), f"Shape mismatch: param {param.shape} vs loaded_weight {loaded_weight.shape}"
-            param.set_value(loaded_weight)
+            param.copy_(loaded_weight, False)
             return
 
         start_idx = self.shard_indices.org_vocab_start_index

fastdeploy/model_executor/models/adapters.py

@@ -22,7 +22,6 @@
 
 from fastdeploy.config import ModelConfig
 from fastdeploy.model_executor.layers.activation import get_act_fn
-from fastdeploy.model_executor.models.interfaces_base import is_pooling_model
 from fastdeploy.transformer_utils.config import get_hf_file_to_dict
 
 _T = TypeVar("_T", bound=type[nn.Layer])
@@ -191,6 +190,8 @@ def as_embedding_model(cls: _T) -> _T:
         please implement your own model if this is not the case.
     """
     # Avoid modifying existing embedding models
+    from fastdeploy.model_executor.models.interfaces_base import is_pooling_model
+
     if is_pooling_model(cls):
         return cls
 

fastdeploy/worker/gpu_model_runner.py

@@ -1321,6 +1321,7 @@ def _dummy_run(
 
             logits = None
             if hasattr(self.model, "is_pooling_model") and self.model.is_pooling_model:
+                # TODO(lizexu123) The preheating the pooling function have not been implemented yet.
                 pass
             else:
                 # 4. Execute spec decode
@@ -1632,9 +1633,9 @@ class at the server level, which is too granular for ModelRunner.
         logits = None
         # 4. Compute logits, Sample
         if hasattr(self.model, "is_pooling_model") and self.model.is_pooling_model:
+            # TODO(lizexu123) The execution of the pooling function have not been implemented yet.
             pass
         else:
-            # 4. Execute spec decode
             logits = self.model.compute_logits(hidden_states)
 
         if not self.speculative_decoding: