[Doc] Update the Pangu Pro MoE tutorials. (#3651)

### What this PR does / why we need it?
Update the Pangu Pro MoE tutorials.
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

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

Signed-off-by: menogrey <1299267905@qq.com>

docs/source/tutorials/multi_npu_moge.md

@@ -51,6 +51,7 @@ vllm serve /path/to/pangu-pro-moe-model \
 --tensor-parallel-size 4 \
 --enable-expert-parallel \
 --trust-remote-code \
+--max_model_len=1024 \
 --enforce-eager
 ```
 
@@ -217,6 +218,7 @@ if __name__ == "__main__":
 
     llm = LLM(model="/path/to/pangu-pro-moe-model",
             tensor_parallel_size=4,
+            enable_expert_parallel=True,
             distributed_executor_backend="mp",
             max_model_len=1024,
             trust_remote_code=True,