[CI]Fix oom of deepseek-eplb nigtly test. (#3884)

### What this PR does / why we need it?
Fix oom of deepseek-eplb nigtly test

- vLLM version: v0.11.0rc3
- vLLM main:
83f478bb19

---------

Signed-off-by: offline0806 <3337230449@qq.com>
Co-authored-by: offline0806 <3337230449@qq.com>

vllm_ascend/eplb/core/eplb_device_transfer_loader.py

@@ -126,7 +126,7 @@ class D2DExpertWeightLoader:
                                                       local_expert_to_replace,
                                                       buffer_tensor_id)
 
-        logger.info(
+        logger.debug(
             f"[EPLB] finished update expert weight for layer: {self.layer_id}")
 
         self.recv_expert_list = []

vllm_ascend/eplb/eplb_updator.py

@@ -77,6 +77,7 @@ class EplbUpdator:
         self.cur_iterations += 1
         if self.cur_iterations == (self.num_iterations_eplb_update + \
                                    self.num_wait_worker_iterations + self.num_moe_layers):
+            logger.info("Finish expert parallel load balancing.")
             if self.expert_map_record_path is not None:
                 self.adaptor._export_tensor_to_file(
                     self.shared_dict["expert_maps"],