RFTConfig¶
Structure Class¶
RFTConfig
dataclass
¶
Configuration settings for reinforcement fine-tuning (RFT), including grader configuration and training hyperparameters.
Attributes¶
grader_config
class-attribute
instance-attribute
¶
grader_config: GraderConfig | None = None
Configuration for the grader that evaluates model responses and provides reward signals during RFT training.
hyper_parameters
class-attribute
instance-attribute
¶
hyper_parameters: RFTHyperParameters | None = None
Hyperparameters that control the reinforcement fine-tuning training process, including learning rate, batch size, and epoch count.