feat: add static embedding metadata to dynamic model listings for providers using OpenAIMixin (#3547)

# What does this PR do?

- remove auto-download of ollama embedding models
- add embedding model metadata to dynamic listing w/ unit test
- add support and tests for allowed_models
- removed inference provider models.py files where dynamic listing is
enabled
- store embedding metadata in embedding_model_metadata field on
inference providers
- make model_entries optional on ModelRegistryHelper and
LiteLLMOpenAIMixin
- make OpenAIMixin a ModelRegistryHelper
- skip base64 embedding test for remote::ollama, always returns floats
- only use OpenAI client for ollama model listing
- remove unused build_model_entry function
- remove unused get_huggingface_repo function


## Test Plan

ci w/ new tests

llama_stack/providers/remote/inference/fireworks/fireworks.py

@@ -54,15 +54,18 @@ from llama_stack.providers.utils.inference.prompt_adapter import (
 )
 
 from .config import FireworksImplConfig
-from .models import MODEL_ENTRIES
 
 logger = get_logger(name=__name__, category="inference::fireworks")
 
 
 class FireworksInferenceAdapter(OpenAIMixin, ModelRegistryHelper, Inference, NeedsRequestProviderData):
+    embedding_model_metadata = {
+        "nomic-ai/nomic-embed-text-v1.5": {"embedding_dimension": 768, "context_length": 8192},
+    }
+
     def __init__(self, config: FireworksImplConfig) -> None:
-        ModelRegistryHelper.__init__(self, MODEL_ENTRIES, config.allowed_models)
         self.config = config
+        self.allowed_models = config.allowed_models
 
     async def initialize(self) -> None:
         pass

llama_stack/providers/remote/inference/fireworks/models.py

@@ -1,70 +0,0 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-#
-# This source code is licensed under the terms described in the LICENSE file in
-# the root directory of this source tree.
-
-from llama_stack.apis.models import ModelType
-from llama_stack.models.llama.sku_types import CoreModelId
-from llama_stack.providers.utils.inference.model_registry import (
-    ProviderModelEntry,
-    build_hf_repo_model_entry,
-)
-
-SAFETY_MODELS_ENTRIES = [
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-guard-3-8b",
-        CoreModelId.llama_guard_3_8b.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-guard-3-11b-vision",
-        CoreModelId.llama_guard_3_11b_vision.value,
-    ),
-]
-
-MODEL_ENTRIES = [
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p1-8b-instruct",
-        CoreModelId.llama3_1_8b_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p1-70b-instruct",
-        CoreModelId.llama3_1_70b_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p1-405b-instruct",
-        CoreModelId.llama3_1_405b_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p2-3b-instruct",
-        CoreModelId.llama3_2_3b_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p2-11b-vision-instruct",
-        CoreModelId.llama3_2_11b_vision_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p2-90b-vision-instruct",
-        CoreModelId.llama3_2_90b_vision_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama-v3p3-70b-instruct",
-        CoreModelId.llama3_3_70b_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama4-scout-instruct-basic",
-        CoreModelId.llama4_scout_17b_16e_instruct.value,
-    ),
-    build_hf_repo_model_entry(
-        "accounts/fireworks/models/llama4-maverick-instruct-basic",
-        CoreModelId.llama4_maverick_17b_128e_instruct.value,
-    ),
-    ProviderModelEntry(
-        provider_model_id="nomic-ai/nomic-embed-text-v1.5",
-        model_type=ModelType.embedding,
-        metadata={
-            "embedding_dimension": 768,
-            "context_length": 8192,
-        },
-    ),
-] + SAFETY_MODELS_ENTRIES