phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-08 19:10:56 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Push registration methods onto the backing providers

Browse source
This commit is contained in:
Ashwin Bharambe 2024-10-05 22:17:06 -07:00 • committed by Ashwin Bharambe
parent 5a7b01d292
commit 4215cc9331
14 changed files with 269 additions and 220 deletions
Download patch file Download diff file Expand all files Collapse all files

10
llama_stack/apis/inference/inference.py
View file

@ -14,6 +14,7 @@ from pydantic import BaseModel, Field
from typing_extensions import Annotated
from llama_models.llama3.api.datatypes import * # noqa: F403
from llama_stack.apis.models import * # noqa: F403
class LogProbConfig(BaseModel):
@ -203,3 +204,12 @@ class Inference(Protocol):
model: str,
contents: List[InterleavedTextMedia],
) -> EmbeddingsResponse: ...
@webmethod(route="/inference/register_model")
async def register_model(self, model: ModelDef) -> None: ...
@webmethod(route="/inference/list_models")
async def list_models(self) -> List[ModelDef]: ...
@webmethod(route="/inference/get_model")
async def get_model(self, identifier: str) -> Optional[ModelDef]: ...