Updating since OpenAIMixin is Pydantic Base Model

llama_stack/providers/remote/inference/runpod/__init__.py

@@ -11,6 +11,6 @@ async def get_adapter_impl(config: RunpodImplConfig, _deps):
     from .runpod import RunpodInferenceAdapter
 
     assert isinstance(config, RunpodImplConfig), f"Unexpected config type: {type(config)}"
-    impl = RunpodInferenceAdapter(config)
+    impl = RunpodInferenceAdapter(config=config)
     await impl.initialize()
     return impl

llama_stack/providers/remote/inference/runpod/runpod.py

@@ -7,35 +7,24 @@
 from typing import Any
 
 from llama_stack.apis.inference import (
-    Inference,
     OpenAIEmbeddingsResponse,
     OpenAIMessageParam,
     OpenAIResponseFormatParam,
 )
 from llama_stack.apis.models import Model
-from llama_stack.providers.utils.inference.model_registry import ModelRegistryHelper
 from llama_stack.providers.utils.inference.openai_mixin import OpenAIMixin
 
 from .config import RunpodImplConfig
 
-MODEL_ENTRIES = []
 
-
+class RunpodInferenceAdapter(OpenAIMixin):
-class RunpodInferenceAdapter(
-    OpenAIMixin,
-    ModelRegistryHelper,
-    Inference,
-):
     """
     Adapter for RunPod's OpenAI-compatible API endpoints.
     Supports VLLM for serverless endpoint self-hosted or public endpoints.
     Can work with any runpod endpoints that support OpenAI-compatible API
     """
 
-    def __init__(self, config: RunpodImplConfig) -> None:
+    config: RunpodImplConfig
-        OpenAIMixin.__init__(self)
-        ModelRegistryHelper.__init__(self, MODEL_ENTRIES)
-        self.config = config
 
     def get_api_key(self) -> str:
         """Get API key for OpenAI client."""
@@ -45,12 +34,6 @@ class RunpodInferenceAdapter(
         """Get base URL for OpenAI client."""
         return self.config.url
 
-    async def initialize(self) -> None:
-        pass
-
-    async def shutdown(self) -> None:
-        pass
-
     async def openai_chat_completion(
         self,
         model: str,