Remove hardcoded list

vllm-project · Oct 1, 2024 · 30e454a · 30e454a
1 parent b4ce5f7
commit 30e454a
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 28 deletions.
diff --git a/vllm/config.py b/vllm/config.py
@@ -397,9 +397,8 @@ def verify_with_parallel_config(
             architectures = getattr(self.hf_config, "architectures", [])
             if not ModelRegistry.is_pp_supported_model(architectures):
                 raise NotImplementedError(
-                    "Pipeline parallelism is only supported for the following "
-                    f"architectures: {ModelRegistry.get_pp_supported_archs()}."
-                )
+                    "Pipeline parallelism is not supported for this model. "
+                    "Supported models implement the `SupportsPP` interface.")
 
             if self.use_async_output_proc:
                 logger.warning("Async output processor is not supported with "

diff --git a/vllm/model_executor/models/__init__.py b/vllm/model_executor/models/__init__.py
@@ -123,27 +123,6 @@
     **_CONDITIONAL_GENERATION_MODELS,
 }
 
-_PP_SUPPORTED_MODELS = [
-    "AquilaForCausalLM",
-    "AquilaModel",
-    "DeepseekV2ForCausalLM",
-    "GPT2LMHeadModel",
-    "InternLM2ForCausalLM",
-    "InternLMForCausalLM",
-    "InternVLChatModel",
-    "JAISLMHeadModel",
-    "LlamaForCausalLM",
-    "LLaMAForCausalLM",
-    "MistralForCausalLM",
-    "MixtralForCausalLM",
-    "NemotronForCausalLM",
-    "Phi3ForCausalLM",
-    "Qwen2ForCausalLM",
-    "Qwen2MoeForCausalLM",
-    "QWenLMHeadModel",
-    "Qwen2VLForConditionalGeneration",
-]
-
 # Architecture -> type.
 # out of tree models
 _OOT_MODELS: Dict[str, Type[nn.Module]] = {}
@@ -297,10 +276,6 @@ def is_pp_supported_model(architectures: Union[str, List[str]]) -> bool:
 
         return any(is_pp(arch) for arch in architectures)
 
-    @staticmethod
-    def get_pp_supported_archs() -> List[str]:
-        return list(_PP_SUPPORTED_MODELS)
-
 
 __all__ = [
     "ModelRegistry",