RFTConfigDict¶
Dictionary Type¶
RFTConfigDict
¶
Bases: TypedDict
Dictionary form of RFTConfig.
Attributes¶
grader_config
instance-attribute
¶
grader_config: GraderConfig | GraderConfigDict | None
Configuration for the grader that evaluates model responses and provides reward signals during RFT training.
hyper_parameters
instance-attribute
¶
hyper_parameters: RFTHyperParameters | RFTHyperParametersDict | None
Hyperparameters that control the reinforcement fine-tuning training process, including learning rate, batch size, and epoch count.