Fixed model name; use routing_key to get model

llama_stack/distribution/resolver.py

@@ -156,6 +156,11 @@ async def instantiate_provider(
         assert isinstance(provider_config, GenericProviderConfig)
         config_type = instantiate_class_type(provider_spec.config_class)
         config = config_type(**provider_config.config)
+
+        if hasattr(provider_config, "routing_key"):
+            routing_key = provider_config.routing_key
+            deps["routing_key"] = routing_key
+
         args = [config, deps]
     elif isinstance(provider_spec, AutoRoutedProviderSpec):
         method = "get_auto_router_impl"

llama_stack/providers/adapters/inference/ollama/__init__.py

@@ -15,5 +15,12 @@ async def get_adapter_impl(config: RemoteProviderConfig, _deps):
     from .ollama import OllamaInferenceAdapter
 
     impl = OllamaInferenceAdapter(config.url)
-    await impl.initialize()
+
+    routing_key = _deps.get("routing_key")
+    if not routing_key:
+        raise ValueError(
+            "Routing key is required for the Ollama adapter but was not found."
+        )
+
+    await impl.initialize(routing_key)
     return impl

llama_stack/providers/adapters/inference/ollama/ollama.py

@@ -23,7 +23,7 @@ from llama_stack.providers.utils.inference.routable import RoutableProviderForMo
 # TODO: Eventually this will move to the llama cli model list command
 # mapping of Model SKUs to ollama models
 OLLAMA_SUPPORTED_SKUS = {
-    "Llama-Guard-3-1B": "xe/llamaguard3:latest",
+    "Llama-Guard-3-8B": "xe/llamaguard3:latest",
     "llama_guard": "xe/llamaguard3:latest",
     "Llama3.1-8B-Instruct": "llama3.1:8b-instruct-fp16",
     "Llama3.1-70B-Instruct": "llama3.1:70b-instruct-fp16",
@@ -40,18 +40,18 @@ class OllamaInferenceAdapter(Inference, RoutableProviderForModels):
         self.url = url
         tokenizer = Tokenizer.get_instance()
         self.formatter = ChatFormat(tokenizer)
-        self.model = "Llama3.2-1B-Instruct"
 
     @property
     def client(self) -> AsyncClient:
         return AsyncClient(host=self.url)
 
-    async def initialize(self) -> None:
+    async def initialize(self, routing_key: str) -> None:
         print("Initializing Ollama, checking connectivity to server...")
         try:
             await self.client.ps()
-            print(f"Connected to Ollama server. Pre-downloading {self.model}...")
-            await self.predownload_models(ollama_model=self.model)
+            ollama_model = self.map_to_provider_model(routing_key)
+            print(f"Connected to Ollama server. Pre-downloading {ollama_model}...")
+            await self.predownload_models(ollama_model)
         except httpx.ConnectError as e:
             raise RuntimeError(
                 "Ollama Server is not running, start it using `ollama serve` in a separate terminal"
@@ -140,8 +140,7 @@ class OllamaInferenceAdapter(Inference, RoutableProviderForModels):
         options = self.get_ollama_chat_options(request)
         ollama_model = self.map_to_provider_model(request.model)
 
-        if ollama_model != self.model:
-            self.predownload_models(ollama_model)
+        self.predownload_models(ollama_model)
 
         if not request.stream:
             r = await self.client.chat(