phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-28 10:32:01 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Address PR comments

Browse source
This commit is contained in:
Jash Gulabrai 2025-04-15 18:04:19 -04:00
parent b8d2e33955
commit 334cd8b34a
8 changed files with 12 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

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

@ -28,7 +28,7 @@ class ModelCandidate(BaseModel):
type: Literal["model"] = "model"
model: str
sampling_params: Optional[SamplingParams] = Field(default_factory=SamplingParams)
sampling_params: SamplingParams
system_message: Optional[SystemMessage] = None

6
llama_stack/providers/remote/eval/nvidia/config.py
View file

@ -14,10 +14,10 @@ class NVIDIAEvalConfig(BaseModel):
Configuration for the NVIDIA NeMo Evaluator microservice endpoint.
Attributes:
evaluator_service_url (str): A base url for accessing the NVIDIA evaluation endpoint, e.g. http://localhost:8000.
evaluator_url (str): A base url for accessing the NVIDIA evaluation endpoint, e.g. http://localhost:8000.
"""
evaluator_service_url: str = Field(
evaluator_url: str = Field(
default_factory=lambda: os.getenv("NVIDIA_EVALUATOR_URL", "http://0.0.0.0:7331"),
description="The url for accessing the evaluator service",
)
@ -25,5 +25,5 @@ class NVIDIAEvalConfig(BaseModel):
@classmethod
def sample_run_config(cls, **kwargs) -> Dict[str, Any]:
return {
"evaluator_service_url": "${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}",
"evaluator_url": "${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}",
}

4
llama_stack/providers/remote/eval/nvidia/eval.py
View file

@ -53,13 +53,13 @@ class NVIDIAEvalImpl(
async def _evaluator_get(self, path):
"""Helper for making GET requests to the evaluator service."""
response = requests.get(url=f"{self.config.evaluator_service_url}{path}")
response = requests.get(url=f"{self.config.evaluator_url}{path}")
response.raise_for_status()
return response.json()
async def _evaluator_post(self, path, data):
"""Helper for making POST requests to the evaluator service."""
response = requests.post(url=f"{self.config.evaluator_service_url}{path}", json=data)
response = requests.post(url=f"{self.config.evaluator_url}{path}", json=data)
response.raise_for_status()
return response.json()

2
llama_stack/templates/nvidia/run-with-safety.yaml
View file

@ -56,7 +56,7 @@ providers:
- provider_id: nvidia
provider_type: remote::nvidia
config:
evaluator_service_url: ${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}
evaluator_url: ${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}
post_training:
- provider_id: nvidia
provider_type: remote::nvidia

2
llama_stack/templates/nvidia/run.yaml
View file

@ -51,7 +51,7 @@ providers:
- provider_id: nvidia
provider_type: remote::nvidia
config:
evaluator_service_url: ${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}
evaluator_url: ${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}
post_training:
- provider_id: nvidia
provider_type: remote::nvidia