Skip to content

llm_tool

llm_tool

LLM tool — delegate a sub-query to an inference engine.

Classes

LLMTool 

LLMTool(
    engine: Optional[InferenceEngine] = None,
    *,
    model: str = "",
)

Bases: BaseTool

Delegate a sub-query to an inference engine for generation.

Source code in src/diapason/tools/llm_tool.py 
19
20
21
22
23
24
25
26
def __init__(
    self,
    engine: Optional[InferenceEngine] = None,
    *,
    model: str = "",
) -> None:
    self._engine = engine
    self._model = model