phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-03 09:53:45 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Refactor list_models method

Browse source
This commit is contained in:
Wojciech-Rebisz 2025-11-04 15:38:18 +01:00
parent 8cae4ddc2a
commit 75e50e87c8
1 changed files with 8 additions and 27 deletions
Download patch file Download diff file Expand all files Collapse all files

35
src/llama_stack/providers/remote/inference/watsonx/watsonx.py
View file

@ -274,42 +274,23 @@ class WatsonXInferenceAdapter(LiteLLMOpenAIMixin):
models = [] models = []
for model_spec in self._get_model_specs(): for model_spec in self._get_model_specs():
functions = [f["id"] for f in model_spec.get("functions", [])] functions = [f["id"] for f in model_spec.get("functions", [])]
# Format: {"embedding_dimension": 1536, "context_length": 8192}
# Example of an embedding model:
# {'model_id': 'ibm/granite-embedding-278m-multilingual',
# 'label': 'granite-embedding-278m-multilingual',
# 'model_limits': {'max_sequence_length': 512, 'embedding_dimension': 768},
# ...
provider_resource_id = f"{self.__provider_id__}/{model_spec['model_id']}" provider_resource_id = f"{self.__provider_id__}/{model_spec['model_id']}"
if "embedding" in functions: if "embedding" in functions:
embedding_dimension = model_spec.get("model_limits", {}).get("embedding_dimension", 0) model_type = ModelType.embedding
context_length = model_spec.get("model_limits", {}).get("max_sequence_length", 0) elif "text_chat" in functions:
embedding_metadata = { model_type = ModelType.llm
"embedding_dimension": embedding_dimension, else:
"context_length": context_length, model_type = None
}
model = Model( if model_type is not None:
identifier=model_spec["model_id"],
provider_resource_id=provider_resource_id,
provider_id=self.__provider_id__,
metadata=embedding_metadata,
model_type=ModelType.embedding,
)
self._model_cache[provider_resource_id] = model
models.append(model)
if "text_chat" in functions:
model = Model( model = Model(
identifier=model_spec["model_id"], identifier=model_spec["model_id"],
provider_resource_id=provider_resource_id, provider_resource_id=provider_resource_id,
provider_id=self.__provider_id__, provider_id=self.__provider_id__,
metadata={}, metadata={},
model_type=ModelType.llm, model_type=model_type,
) )
# In theory, I guess it is possible that a model could be both an embedding model and a text chat model.
# In that case, the cache will record the generator Model object, and the list which we return will have
# both the generator Model object and the text chat Model object. That's fine because the cache is
# only used for check_model_availability() anyway.
self._model_cache[provider_resource_id] = model self._model_cache[provider_resource_id] = model
models.append(model) models.append(model)
return models return models