FineTuneHyperparams.md

OpenapiClient::FineTuneHyperparams

Properties

Name	Type	Description	Notes
batch_size	Integer	The batch size to use for training. The batch size is the number of training examples used to train a single forward and backward pass.
classification_n_classes	Integer	The number of classes to use for computing classification metrics.	[optional]
classification_positive_class	String	The positive class to use for computing classification metrics.	[optional]
compute_classification_metrics	Boolean	The classification metrics to compute using the validation dataset at the end of every epoch.	[optional]
learning_rate_multiplier	Float	The learning rate multiplier to use for training.
n_epochs	Integer	The number of epochs to train the model for. An epoch refers to one full cycle through the training dataset.
prompt_loss_weight	Float	The weight to use for loss on the prompt tokens.

Example

require 'openapi_client'

instance = OpenapiClient::FineTuneHyperparams.new(
  batch_size: null,
  classification_n_classes: null,
  classification_positive_class: null,
  compute_classification_metrics: null,
  learning_rate_multiplier: null,
  n_epochs: null,
  prompt_loss_weight: null
)