phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2026-01-02 08:24:38 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

change to dev, fix issues with test

Browse source
This commit is contained in:
Honglin Cao 2025-03-05 19:17:45 -05:00
parent e31a52b26e
commit 941d5f1b18
2 changed files with 8 additions and 28 deletions
Download patch file Download diff file Expand all files Collapse all files

28
llama_stack/providers/remote/inference/centml/centml.py
View file

@ -256,32 +256,6 @@ class CentMLInferenceAdapter(
logcat.debug("inference", f"params to centml: {params}")
return params
def _build_options(
self,
sampling_params: Optional[SamplingParams],
fmt: Optional[ResponseFormat],
) -> dict:
"""
Build temperature, max_tokens, top_p, etc., plus any response format data.
"""
options = get_sampling_options(sampling_params)
options.setdefault("max_tokens", 512)
if fmt:
if fmt.type == ResponseFormatType.json_schema.value:
options["response_format"] = {
"type": "json_object",
"schema": fmt.json_schema,
}
elif fmt.type == ResponseFormatType.grammar.value:
raise NotImplementedError(
"Grammar response format not supported yet"
)
else:
raise ValueError(f"Unknown response format {fmt.type}")
return options
def _build_options(
self,
sampling_params: Optional[SamplingParams],
@ -302,7 +276,7 @@ class CentMLInferenceAdapter(
raise ValueError(f"Unknown response format {fmt.type}")
if logprobs and logprobs.top_k:
options["logprobs"] = 1
options["logprobs"] = logprobs.top_k
return options