refactor: replace default all-MiniLM-L6-v2 embedding model by nomic-embed-text-v1.5 in Llama Stack (#3183)

# What does this PR do?
<!-- Provide a short summary of what this PR does and why. Link to
relevant issues if applicable. -->
The purpose of this PR is to replace the Llama Stack's default embedding
model by nomic-embed-text-v1.5.

These are the key reasons why Llama Stack community decided to switch
from all-MiniLM-L6-v2 to nomic-embed-text-v1.5:
1. The training data for
[all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2#training-data)
includes a lot of data sets with various licensing terms, so it is
tricky to know when/whether it is appropriate to use this model for
commercial applications.
2. The model is not particularly competitive on major benchmarks. For
example, if you look at the [MTEB
Leaderboard](https://huggingface.co/spaces/mteb/leaderboard) and click
on Miscellaneous/BEIR to see English information retrieval accuracy, you
see that the top of the leaderboard is dominated by enormous models but
also that there are many, many models of relatively modest size whith
much higher Retrieval scores. If you want to look closely at the data, I
recommend clicking "Download Table" because it is easier to browse that
way.

More discussion info can be founded
[here](https://github.com/llamastack/llama-stack/issues/2418)

<!-- If resolving an issue, uncomment and update the line below -->
<!-- Closes #[issue-number] -->
Closes #2418 

## Test Plan
<!-- Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.* -->
1. Run `./scripts/unit-tests.sh`
2. Integration tests via CI wokrflow

---------

Signed-off-by: Sébastien Han <seb@redhat.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Francisco Arceo <arceofrancisco@gmail.com>
Co-authored-by: Sébastien Han <seb@redhat.com>

llama_stack/providers/inline/inference/sentence_transformers/sentence_transformers.py

@@ -54,11 +54,11 @@ class SentenceTransformersInferenceImpl(
     async def list_models(self) -> list[Model] | None:
         return [
             Model(
-                identifier="all-MiniLM-L6-v2",
-                provider_resource_id="all-MiniLM-L6-v2",
+                identifier="nomic-ai/nomic-embed-text-v1.5",
+                provider_resource_id="nomic-ai/nomic-embed-text-v1.5",
                 provider_id=self.__provider_id__,
                 metadata={
-                    "embedding_dimension": 384,
+                    "embedding_dimension": 768,
                 },
                 model_type=ModelType.embedding,
             ),

llama_stack/providers/registry/inference.py

@@ -43,6 +43,12 @@ def available_providers() -> list[ProviderSpec]:
             pip_packages=[
                 "torch torchvision torchao>=0.12.0 --extra-index-url https://download.pytorch.org/whl/cpu",
                 "sentence-transformers --no-deps",
+                # required by some SentenceTransformers architectures for tensor rearrange/merge ops
+                "einops",
+                # fast HF tokenization backend used by SentenceTransformers models
+                "tokenizers",
+                # safe and fast file format for storing and loading tensors
+                "safetensors",
             ],
             module="llama_stack.providers.inline.inference.sentence_transformers",
             config_class="llama_stack.providers.inline.inference.sentence_transformers.config.SentenceTransformersInferenceConfig",

llama_stack/providers/utils/inference/embedding_mixin.py

@@ -83,7 +83,7 @@ class SentenceTransformerEmbeddingMixin:
         def _load_model():
             from sentence_transformers import SentenceTransformer
 
-            return SentenceTransformer(model)
+            return SentenceTransformer(model, trust_remote_code=True)
 
         loaded_model = await asyncio.to_thread(_load_model)
         EMBEDDING_MODELS[model] = loaded_model

llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -352,17 +352,16 @@ class OpenAIVectorStoreMixin(ABC):
         extra = params.model_extra or {}
         provider_vector_db_id = extra.get("provider_vector_db_id")
         embedding_model = extra.get("embedding_model")
-        embedding_dimension = extra.get("embedding_dimension", 384)
+        embedding_dimension = extra.get("embedding_dimension", 768)
         # use provider_id set by router; fallback to provider's own ID when used directly via --stack-config
         provider_id = extra.get("provider_id") or getattr(self, "__provider_id__", None)
-
         # Derive the canonical vector_db_id (allow override, else generate)
         vector_db_id = provider_vector_db_id or generate_object_id("vector_store", lambda: f"vs_{uuid.uuid4()}")
 
         if embedding_model is None:
             raise ValueError("Embedding model is required")
 
-        # Embedding dimension is required (defaulted to 384 if not provided)
+        # Embedding dimension is required (defaulted to 768 if not provided)
         if embedding_dimension is None:
             raise ValueError("Embedding dimension is required")