model update and delete for provider

llama_stack/providers/datatypes.py

@@ -45,6 +45,10 @@ class Api(Enum):
 class ModelsProtocolPrivate(Protocol):
     async def register_model(self, model: Model) -> None: ...
 
+    async def update_model(self, model: Model) -> None: ...
+
+    async def unregister_model(self, model_id: str) -> None: ...
+
 
 class ShieldsProtocolPrivate(Protocol):
     async def register_shield(self, shield: Shield) -> None: ...

llama_stack/providers/inline/inference/meta_reference/inference.py

@@ -71,6 +71,12 @@ class MetaReferenceInferenceImpl(Inference, ModelRegistryHelper, ModelsProtocolP
                 f"Model mismatch: {request.model} != {self.model.descriptor()}"
             )
 
+    async def update_model(self, model: Model) -> None:
+        pass
+
+    async def unregister_model(self, model_id: str) -> None:
+        pass
+
     async def completion(
         self,
         model_id: str,

llama_stack/providers/inline/inference/vllm/vllm.py

@@ -108,6 +108,12 @@ class VLLMInferenceImpl(Inference, ModelsProtocolPrivate):
 
         return VLLMSamplingParams(**kwargs)
 
+    async def update_model(self, model: Model) -> None:
+        pass
+
+    async def unregister_model(self, model_id: str) -> None:
+        pass
+
     async def completion(
         self,
         model_id: str,

llama_stack/providers/remote/inference/ollama/ollama.py

@@ -93,6 +93,12 @@ class OllamaInferenceAdapter(Inference, ModelsProtocolPrivate):
     async def shutdown(self) -> None:
         pass
 
+    async def update_model(self, model: Model) -> None:
+        pass
+
+    async def unregister_model(self, model_id: str) -> None:
+        pass
+
     async def completion(
         self,
         model_id: str,

llama_stack/providers/remote/inference/tgi/tgi.py

@@ -69,6 +69,12 @@ class _HfAdapter(Inference, ModelsProtocolPrivate):
     async def shutdown(self) -> None:
         pass
 
+    async def update_model(self, model: Model) -> None:
+        pass
+
+    async def unregister_model(self, model_id: str) -> None:
+        pass
+
     async def completion(
         self,
         model: str,

llama_stack/providers/remote/inference/vllm/vllm.py

@@ -58,6 +58,12 @@ class VLLMInferenceAdapter(Inference, ModelsProtocolPrivate):
     async def shutdown(self) -> None:
         pass
 
+    async def update_model(self, model: Model) -> None:
+        pass
+
+    async def unregister_model(self, model_id: str) -> None:
+        pass
+
     async def completion(
         self,
         model_id: str,