phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-09 13:14:39 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

test

Browse source 
# What does this PR do?


## Test Plan
This commit is contained in:
Eric Huang 2025-10-08 13:23:21 -07:00
parent 96886afaca
commit 1e891489a8
13 changed files with 170 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

2
llama_stack/apis/inference/inference.py
View file

@ -1106,6 +1106,7 @@ class InferenceProvider(Protocol):
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
**kwargs: Any,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
"""Create chat completions.
@ -1134,6 +1135,7 @@ class InferenceProvider(Protocol):
:param top_logprobs: (Optional) The top log probabilities to use.
:param top_p: (Optional) The top p to use.
:param user: (Optional) The user to use.
:param kwargs: (Optional) Additional provider-specific parameters to pass through as extra_body (e.g., chat_template_kwargs for vLLM).
:returns: An OpenAIChatCompletion.
"""
...

2
llama_stack/core/routers/inference.py
View file

@ -277,6 +277,7 @@ class InferenceRouter(Inference):
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
**kwargs: Any,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
logger.debug(
f"InferenceRouter.openai_chat_completion: {model=}, {stream=}, {messages=}",
@ -323,6 +324,7 @@ class InferenceRouter(Inference):
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
**kwargs,
)
provider = await self.routing_table.get_provider_impl(model_obj.identifier)
if stream:

2
llama_stack/providers/remote/inference/passthrough/passthrough.py
View file

@ -153,6 +153,7 @@ class PassthroughInferenceAdapter(Inference):
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
**kwargs: Any,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
client = self._get_client()
model_obj = await self.model_store.get_model(model)
@ -181,6 +182,7 @@ class PassthroughInferenceAdapter(Inference):
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
**kwargs,
)
return await client.inference.openai_chat_completion(**params)

2
llama_stack/providers/remote/inference/runpod/runpod.py
View file

@ -57,6 +57,7 @@ class RunpodInferenceAdapter(OpenAIMixin):
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
**kwargs: Any,
):
"""Override to add RunPod-specific stream_options requirement."""
if stream and not stream_options:
@ -86,4 +87,5 @@ class RunpodInferenceAdapter(OpenAIMixin):
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
**kwargs,
)

2
llama_stack/providers/remote/inference/vllm/vllm.py
View file

@ -102,6 +102,7 @@ class VLLMInferenceAdapter(OpenAIMixin):
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
**kwargs: Any,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
max_tokens = max_tokens or self.config.max_tokens
@ -136,4 +137,5 @@ class VLLMInferenceAdapter(OpenAIMixin):
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
**kwargs,
)

6
llama_stack/providers/utils/inference/openai_mixin.py
View file

@ -313,6 +313,7 @@ class OpenAIMixin(NeedsRequestProviderData, ABC, BaseModel):
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
**kwargs: Any,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
"""
Direct OpenAI chat completion API call.
@ -361,7 +362,10 @@ class OpenAIMixin(NeedsRequestProviderData, ABC, BaseModel):
user=user,
)
resp = await self.client.chat.completions.create(**params)
# Pass any additional provider-specific parameters as extra_body
extra_body = kwargs if kwargs else {}
resp = await self.client.chat.completions.create(**params, extra_body=extra_body)
return await self._maybe_overwrite_id(resp, stream) # type: ignore[no-any-return]