[edge-infer] Spark2_5: AutoWeightsLoader lost skip_prefixes in vLLM 0.29.0

The 0.29.0 loader de-duplicates tied parameters itself, so the explicit
lm_head. skip for tie_word_embeddings is no longer needed (and its keyword
no longer exists: engine init failed with TypeError on the first 0.29.0
parity run).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

vllm_omni/model_executor/models/spark2_5/spark2_5.py

@@ -505,10 +505,10 @@ class Spark2_5ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
         return self.logits_processor(self.lm_head, hidden_states)
 
     def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
-        loader = AutoWeightsLoader(
-            self,
-            skip_prefixes=(["lm_head."] if self.config.tie_word_embeddings else None),
-        )
+        # vLLM 0.29.0 dropped AutoWeightsLoader's ``skip_prefixes``: the loader now
+        # detects tied parameters itself (``_get_tied_embedding_params``) and loads
+        # the first qualname only, so the tied ``lm_head.`` needs no explicit skip.
+        loader = AutoWeightsLoader(self)
         return loader.load_weights(
             _split_fused_qkv(weights, self.config), mapper=self.hf_to_vllm_mapper
         )