bugfix for mtp when running torchair in a2 (#3354)

### What this PR does / why we need it?
when ops torchair_fused_experts_with_mc2 is called, we need pass a tp
group, but now it only pass when quantized scenario, we need also pass
it when unquantized.

- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

Signed-off-by: zouyida2052 <zouyida2002@gmail.com>

vllm_ascend/torchair/ops/torchair_fused_moe.py

@@ -899,6 +899,7 @@ class TorchairAscendUnquantizedFusedMoEMethod(UnquantizedFusedMoEMethod):
                 expert_map=expert_map,
                 moe_all_to_all_group_name=self.moe_all_to_all_group_name,
                 shared_experts=shared_experts,
+                is_torchair=self.torchair_graph_enabled,
                 mc2_mask=kwargs.get("mc2_mask", None))
         elif fused_moe_state in [
                 FusedMoEState.AllGather, FusedMoEState.NaiveMulticast