ReinforcementTuningSpec

@Serializable
data class ReinforcementTuningSpec(val compositeRewardConfig: CompositeReinforcementTuningRewardConfig? = null, val trainingDatasetUri: String? = null, val validationDatasetUri: String? = null, val hyperParameters: ReinforcementTuningHyperParameters? = null, val singleRewardConfig: SingleReinforcementTuningRewardConfig? = null)

Reinforcement tuning spec for tuning.

Constructors

Link copied to clipboard
constructor(compositeRewardConfig: CompositeReinforcementTuningRewardConfig? = null, trainingDatasetUri: String? = null, validationDatasetUri: String? = null, hyperParameters: ReinforcementTuningHyperParameters? = null, singleRewardConfig: SingleReinforcementTuningRewardConfig? = null)

Properties

Link copied to clipboard

Composite reward function configuration for reinforcement tuning.

Link copied to clipboard

Additional hyper-parameters to use during tuning.

Link copied to clipboard

Single reward function configuration for reinforcement tuning.

Link copied to clipboard

Cloud Storage path to the file containing training dataset for tuning. The dataset must be formatted as a JSONL file.

Link copied to clipboard

Cloud Storage path to the file containing validation dataset for tuning. The dataset must be formatted as a JSONL file.