OptionaladapterOptional. Adapter size for preference optimization.
OptionalbetaOptional. Weight for KL Divergence regularization.
OptionalepochOptional. Number of complete passes the model makes over the entire training dataset during training.
OptionallearningOptional. Multiplier for adjusting the default learning rate.
Hyperparameters for Preference Optimization. This data type is not supported in Gemini API.