feat(model): New metrics param to model API

pilot/model/base.py

@@ -54,6 +54,8 @@ class ModelOutput:
     model_context: Dict = None
     finish_reason: str = None
     usage: Dict[str, Any] = None
+    metrics: Dict[str, Any] = None
+    """Some metrics for model inference"""
 
     def to_dict(self) -> Dict:
         return asdict(self)

pilot/model/cluster/base.py

@@ -21,6 +21,8 @@ class PromptRequest(BaseModel):
     context_len: int = None
     echo: bool = True
     span_id: str = None
+    metrics: bool = False
+    """Whether to return metrics of inference"""
 
 
 class EmbeddingsRequest(BaseModel):