PreferenceOptimizationHyperParameters

@Serializable

data class PreferenceOptimizationHyperParameters(val adapterSize: AdapterSize? = null, val beta: Double? = null, val epochCount: Long? = null, val learningRateMultiplier: Double? = null)

Hyperparameters for Preference Optimization. This data type is not supported in Gemini API.

Constructors

constructor(adapterSize: AdapterSize? = null, beta: Double? = null, epochCount: Long? = null, learningRateMultiplier: Double? = null)

Optional. Adapter size for preference optimization.

val beta: Double? = null

Optional. Weight for KL Divergence regularization.

Optional. Number of complete passes the model makes over the entire training dataset during training.

Optional. Multiplier for adjusting the default learning rate.