fix model config

Signed-off-by: MengqingCao <cmq0113@163.com>

Author: MengqingCao

tests/singlecard/spec_decode/e2e/test_v1_spec_decode.py

@@ -75,7 +75,7 @@ def test_ngram_correctness(
     with monkeypatch.context() as m:
         m.setenv("VLLM_USE_V1", "1")
 
-        ref_llm = LLM(model=model_name, max_model_len=1024)
+        ref_llm = LLM(model=model_name, max_model_len=1024, enforce_eager=True)
         ref_outputs = ref_llm.chat(test_prompts, sampling_config)
         del ref_llm
 
@@ -88,6 +88,7 @@ def test_ngram_correctness(
                 "num_speculative_tokens": 3,
             },
             max_model_len=1024,
+            enforce_eager=True,
         )
         spec_outputs = spec_llm.chat(test_prompts, sampling_config)
         matches = 0
@@ -138,6 +139,7 @@ def test_eagle_correctness(
                 "max_model_len": 2048,
             },
             max_model_len=2048,
+            enforce_eager=True,
         )
         spec_outputs = spec_llm.chat(test_prompts, sampling_config)
         matches = 0

vllm_ascend/platform.py

@@ -141,7 +141,7 @@ def check_and_update_config(cls, vllm_config: VllmConfig) -> None:
                     "NPU graph mode is still experimental and not supported for V1 without mla currently, "
                     "it has been disabled automatically.")
                 vllm_config.additional_config["enable_graph_mode"] = False
-        elif envs.VLLM_USE_V1 and not enforce_eager:
+        elif envs.VLLM_USE_V1 and model_config is not None and not enforce_eager:
             model_type = model_config.hf_config.model_type
             if "deepseek" in model_type:
                 raise NotImplementedError(