phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-08-10 04:08:31 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

push

Browse source
This commit is contained in:
Xi Yan 2025-02-14 17:35:47 -08:00
parent 34a15014dd
commit 7da7f24504
2 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

2
llama_stack/distribution/routers/routing_tables.py
View file

@ -236,8 +236,6 @@ class ModelsRoutingTable(CommonRoutingTableImpl, Models):
metadata = {} metadata = {}
if model_type is None: if model_type is None:
model_type = ModelType.llm model_type = ModelType.llm
if "embedding_dimension" not in metadata and model_type == ModelType.embedding:
raise ValueError("Embedding model must have an embedding dimension in its metadata")
model = Model( model = Model(
identifier=model_id, identifier=model_id,
provider_resource_id=provider_model_id, provider_resource_id=provider_model_id,

3
llama_stack/providers/inline/inference/sentence_transformers/sentence_transformers.py
View file

@ -19,6 +19,7 @@ from llama_stack.apis.inference import (
ToolDefinition, ToolDefinition,
ToolPromptFormat, ToolPromptFormat,
) )
from llama_stack.apis.models import ModelType
from llama_stack.providers.datatypes import Model, ModelsProtocolPrivate from llama_stack.providers.datatypes import Model, ModelsProtocolPrivate
from llama_stack.providers.utils.inference.embedding_mixin import ( from llama_stack.providers.utils.inference.embedding_mixin import (
SentenceTransformerEmbeddingMixin, SentenceTransformerEmbeddingMixin,
@ -44,6 +45,8 @@ class SentenceTransformersInferenceImpl(
pass pass
async def register_model(self, model: Model) -> None: async def register_model(self, model: Model) -> None:
if "embedding_dimension" not in model.metadata and model.model_type == ModelType.embedding:
raise ValueError("Embedding model must have an embedding dimension in its metadata")
_ = self._load_sentence_transformer_model(model.provider_resource_id) _ = self._load_sentence_transformer_model(model.provider_resource_id)
return model return model