Revert "drop ascend scheduler" (#4580)

Reverts vllm-project/vllm-ascend#4498
- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2

tests/e2e/nightly/models/test_qwen2_5_vl_32b.py

@@ -85,8 +85,9 @@ async def test_models(model: str, tp_size: int) -> None:
         str(tp_size), "--port",
         str(port), "--max-model-len", "30000", "--max-num-batched-tokens",
         "40000", "--max-num-seqs", "400", "--trust-remote-code",
-        "--gpu-memory-utilization", "0.8", "--compilation_config",
-        '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+        "--gpu-memory-utilization", "0.8", "--additional-config",
+        '{"ascend_scheduler_config":{"enabled":false}}',
+        "--compilation_config", '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
     ]
     request_keyword_args: dict[str, Any] = {
         **api_keyword_args,