class

OpenAI::FineTuneDPOHyperparameters

Inherits JSON::Serializable < Reference < Object

The hyperparameters used for the DPO fine-tuning job.

Constructors

new(beta : FineTuneDPOHyperparametersBeta | Nil = nil, batch_size : FineTuneDPOHyperparametersBatchSize | Nil = nil, learning_rate_multiplier : FineTuneDPOHyperparametersLearningRateMultiplier | Nil = nil, n_epochs : FineTuneDPOHyperparametersNEpochs | Nil = nil)

Source

new(*, __pull_for_json_serializable pull : JSON::PullParser)

Source

Instance methods

batch_size

Number of examples in each batch. A larger batch size means that model parameters are updated less frequently, but with lower variance.

Source

batch_size=(batch_size : FineTuneDPOHyperparametersBatchSize | Nil)

Number of examples in each batch. A larger batch size means that model parameters are updated less frequently, but with lower variance.

Source

beta

The beta value for the DPO method. A higher beta value will increase the weight of the penalty between the policy and reference model.

Source

beta=(beta : FineTuneDPOHyperparametersBeta | Nil)

The beta value for the DPO method. A higher beta value will increase the weight of the penalty between the policy and reference model.

Source

learning_rate_multiplier

Scaling factor for the learning rate. A smaller learning rate may be useful to avoid overfitting.

Source

learning_rate_multiplier=(learning_rate_multiplier : FineTuneDPOHyperparametersLearningRateMultiplier | Nil)

Scaling factor for the learning rate. A smaller learning rate may be useful to avoid overfitting.

Source

n_epochs

The number of epochs to train the model for. An epoch refers to one full cycle through the training dataset.

Source

n_epochs=(n_epochs : FineTuneDPOHyperparametersNEpochs | Nil)

The number of epochs to train the model for. An epoch refers to one full cycle through the training dataset.

Source