fix ge check

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

Author: MengqingCao

vllm_ascend/platform.py

@@ -23,7 +23,6 @@
 import vllm.envs as envs
 from vllm.logger import logger
 from vllm.platforms import Platform, PlatformEnum
-from vllm.utils import supports_dynamo
 
 from vllm_ascend.utils import ASCEND_QUATIZATION_METHOD, update_aclgraph_sizes
 
@@ -131,16 +130,27 @@ def check_and_update_config(cls, vllm_config: VllmConfig) -> None:
         if vllm_config.additional_config is not None:
             enable_graph_mode = vllm_config.additional_config.get(
                 "enable_graph_mode", False)
-            if enable_graph_mode and not supports_dynamo():
-                logger.warning(
-                    "enable_graph_mode is not supported because the version of torch is too low, forcing close enable_graph_mode"
-                )
-                vllm_config.additional_config["enable_graph_mode"] = False
-            if enable_graph_mode and envs.VLLM_USE_V1 and envs.VLLM_MLA_DISABLE:
-                logger.warning(
-                    "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
+            if enable_graph_mode:
+                if enforce_eager:
+                    raise RuntimeError(
+                        f"Expect enable_graph_mode != enforce_eager. However got enable_graph_mode={enable_graph_mode}"
+                        f" and enforce_eager={enforce_eager}. Please set `enforce_eager=False` if you attempt"
+                        " to enable NPU graph mode.")
+                elif envs.VLLM_USE_V1 and envs.VLLM_MLA_DISABLE:
+                    logger.warning(
+                        "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
+                if model_config:
+                    model_type = model_config.hf_config.model_type
+                    if "deepseek" not in model_type:
+                        raise NotImplementedError(
+                            "NPU Graph only supports deepseek. Please "
+                            "disable it by additional_config={'enable_graph_mode': True} "
+                            "to serve deepseek models with NPU graph mode on vllm-ascend with V0 engine."
+                            " Or set `enforce_eager=True` to use eager mode on V1 engine."
+                        )
+
         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: