Inference to use provider resource id to register and validate (#428)

This PR changes the way model id gets translated to the final model name
that gets passed through the provider.
Major changes include:
1) Providers are responsible for registering an object and as part of
the registration returning the object with the correct provider specific
name of the model provider_resource_id
2) To help with the common look ups different names a new ModelLookup
class is created.



Tested all inference providers including together, fireworks, vllm,
ollama, meta reference and bedrock

llama_stack/providers/tests/inference/fixtures.py

@@ -49,7 +49,7 @@ def inference_meta_reference(inference_model) -> ProviderFixture:
         providers=[
             Provider(
                 provider_id=f"meta-reference-{i}",
-                provider_type="meta-reference",
+                provider_type="inline::meta-reference",
                 config=MetaReferenceInferenceConfig(
                     model=m,
                     max_seq_len=4096,
@@ -142,6 +142,31 @@ def inference_bedrock() -> ProviderFixture:
     )
 
 
+def get_model_short_name(model_name: str) -> str:
+    """Convert model name to a short test identifier.
+
+    Args:
+        model_name: Full model name like "Llama3.1-8B-Instruct"
+
+    Returns:
+        Short name like "llama_8b" suitable for test markers
+    """
+    model_name = model_name.lower()
+    if "vision" in model_name:
+        return "llama_vision"
+    elif "3b" in model_name:
+        return "llama_3b"
+    elif "8b" in model_name:
+        return "llama_8b"
+    else:
+        return model_name.replace(".", "_").replace("-", "_")
+
+
+@pytest.fixture(scope="session")
+def model_id(inference_model) -> str:
+    return get_model_short_name(inference_model)
+
+
 INFERENCE_FIXTURES = [
     "meta_reference",
     "ollama",

llama_stack/providers/tests/inference/test_text_inference.py

@@ -96,7 +96,7 @@ class TestInference:
         response = await inference_impl.completion(
             content="Micheael Jordan is born in ",
             stream=False,
-            model=inference_model,
+            model_id=inference_model,
             sampling_params=SamplingParams(
                 max_tokens=50,
             ),
@@ -110,7 +110,7 @@ class TestInference:
             async for r in await inference_impl.completion(
                 content="Roses are red,",
                 stream=True,
-                model=inference_model,
+                model_id=inference_model,
                 sampling_params=SamplingParams(
                     max_tokens=50,
                 ),
@@ -171,7 +171,7 @@ class TestInference:
     ):
         inference_impl, _ = inference_stack
         response = await inference_impl.chat_completion(
-            model=inference_model,
+            model_id=inference_model,
             messages=sample_messages,
             stream=False,
             **common_params,
@@ -204,7 +204,7 @@ class TestInference:
             num_seasons_in_nba: int
 
         response = await inference_impl.chat_completion(
-            model=inference_model,
+            model_id=inference_model,
             messages=[
                 SystemMessage(content="You are a helpful assistant."),
                 UserMessage(content="Please give me information about Michael Jordan."),
@@ -227,7 +227,7 @@ class TestInference:
         assert answer.num_seasons_in_nba == 15
 
         response = await inference_impl.chat_completion(
-            model=inference_model,
+            model_id=inference_model,
             messages=[
                 SystemMessage(content="You are a helpful assistant."),
                 UserMessage(content="Please give me information about Michael Jordan."),
@@ -250,7 +250,7 @@ class TestInference:
         response = [
             r
             async for r in await inference_impl.chat_completion(
-                model=inference_model,
+                model_id=inference_model,
                 messages=sample_messages,
                 stream=True,
                 **common_params,
@@ -286,7 +286,7 @@ class TestInference:
         ]
 
         response = await inference_impl.chat_completion(
-            model=inference_model,
+            model_id=inference_model,
             messages=messages,
             tools=[sample_tool_definition],
             stream=False,
@@ -327,7 +327,7 @@ class TestInference:
         response = [
             r
             async for r in await inference_impl.chat_completion(
-                model=inference_model,
+                model_id=inference_model,
                 messages=messages,
                 tools=[sample_tool_definition],
                 stream=True,