ReasoningModelConfig#
- class nemo_microservices.types.shared_params.ReasoningModelConfig
Bases:
TypedDict- end_token: str
The end token used for reasoning traces.
- remove_reasoning_traces: bool
For reasoning models (e.g.
DeepSeek-r1), if the output parser should remove reasoning traces.
- remove_thinking_traces: bool
- start_token: str
The start token used for reasoning traces.