fix aisbench

Signed-off-by: jiangyunfan1 <jiangyunfan1@h-partners.com>

Author: jiangyunfan1

tests/e2e/nightly/features/test_mtpx_deepseek_r1_0528_w8a8.py

@@ -42,7 +42,7 @@
     "case_type": "accuracy",
     "dataset_path": "vllm-ascend/gsm8k-lite",
     "request_conf": "vllm_api_general_chat",
-    "dataset_conf": "gsm8k/gsm8k_gen_0_shot_noncot_chat_prompt",
+    "dataset_conf": "gsm8k/gsm8k_gen_0_shot_cot_chat_prompt",
     "max_out_len": 32768,
     "batch_size": 32,
     "baseline": 95,
@@ -60,8 +60,6 @@
     "threshold": 7
 }]
 
-aisbench_case_dict = {"mtp2": aisbench_gsm8k, "mtp3": aisbench_aime}
-
 
 @pytest.mark.asyncio
 @pytest.mark.parametrize("model", MODELS)
@@ -114,6 +112,7 @@ async def test_models(model: str, mode: str) -> None:
              json.dumps(speculative_config)])
         server_args.extend(["--gpu-memory-utilization", "0.92"])
         additional_config["torchair_graph_config"] = {"enabled": True}
+        aisbench_cases = aisbench_gsm8k
     if mode == "mtp3":
         env_dict["HCCL_OP_EXPANSION_MODE"] = "AIV"
         server_args.extend(["--max-num-batched-tokens", "2048"])
@@ -126,6 +125,7 @@ async def test_models(model: str, mode: str) -> None:
             ["--compilation-config",
              json.dumps(compilation_config)])
         additional_config["torchair_graph_config"] = {"enabled": False}
+        aisbench_cases = aisbench_aime
     server_args.extend(["--additional-config", json.dumps(additional_config)])
     request_keyword_args: dict[str, Any] = {
         **api_keyword_args,
@@ -145,7 +145,6 @@ async def test_models(model: str, mode: str) -> None:
         assert choices[0].text, "empty response"
         print(choices)
         # aisbench test
-        aisbench_cases = aisbench_case_dict[mode]
         run_aisbench_cases(model,
                            port,
                            aisbench_cases,