Class: OpenAI::Models::FineTuning::ReinforcementHyperparameters

Inherits:
Internal::Type::BaseModel show all
Defined in:
lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb

Defined Under Namespace

Modules: BatchSize, ComputeMultiplier, EvalInterval, EvalSamples, LearningRateMultiplier, NEpochs, ReasoningEffort

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from Internal::Type::BaseModel

==, #==, #[], coerce, #deconstruct_keys, #deep_to_h, dump, fields, hash, #hash, inherited, inspect, #inspect, known_fields, optional, recursively_to_h, required, #to_h, #to_json, #to_s, to_sorbet_type, #to_yaml

Methods included from Internal::Type::Converter

#coerce, coerce, #dump, dump, #inspect, inspect, type_info

Methods included from Internal::Util::SorbetRuntimeSupport

#const_missing, #define_sorbet_constant!, #sorbet_constant_defined?, #to_sorbet_type, to_sorbet_type

Constructor Details

#initialize(batch_size: nil, compute_multiplier: nil, eval_interval: nil, eval_samples: nil, learning_rate_multiplier: nil, n_epochs: nil, reasoning_effort: nil) ⇒ Object

Some parameter documentations has been truncated, see OpenAI::Models::FineTuning::ReinforcementHyperparameters for more details.

The hyperparameters used for the reinforcement fine-tuning job.

Parameters:

  • batch_size (Symbol, :auto, Integer) (defaults to: nil) —

    Number of examples in each batch. A larger batch size means that model parameter

  • compute_multiplier (Symbol, :auto, Float) (defaults to: nil) —

    Multiplier on amount of compute used for exploring search space during training.

  • eval_interval (Symbol, :auto, Integer) (defaults to: nil) —

    The number of training steps between evaluation runs.

  • eval_samples (Symbol, :auto, Integer) (defaults to: nil) —

    Number of evaluation samples to generate per training step.

  • learning_rate_multiplier (Symbol, :auto, Float) (defaults to: nil) —

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

  • n_epochs (Symbol, :auto, Integer) (defaults to: nil) —

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

  • reasoning_effort (Symbol, OpenAI::Models::FineTuning::ReinforcementHyperparameters::ReasoningEffort) (defaults to: nil) —

    Level of reasoning effort.




# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 57

Instance Attribute Details

#batch_size ⇒ Symbol, ...

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

Returns:

  • (Symbol, :auto, Integer, nil) 


12
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 12

optional :batch_size, union: -> { OpenAI::FineTuning::ReinforcementHyperparameters::BatchSize }

#compute_multiplier ⇒ Symbol, ...

Multiplier on amount of compute used for exploring search space during training.

Returns:

  • (Symbol, :auto, Float, nil) 


18
19
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 18

optional :compute_multiplier,
union: -> { OpenAI::FineTuning::ReinforcementHyperparameters::ComputeMultiplier }

#eval_interval ⇒ Symbol, ...

The number of training steps between evaluation runs.

Returns:

  • (Symbol, :auto, Integer, nil) 


25
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 25

optional :eval_interval, union: -> { OpenAI::FineTuning::ReinforcementHyperparameters::EvalInterval }

#eval_samples ⇒ Symbol, ...

Number of evaluation samples to generate per training step.

Returns:

  • (Symbol, :auto, Integer, nil) 


31
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 31

optional :eval_samples, union: -> { OpenAI::FineTuning::ReinforcementHyperparameters::EvalSamples }

#learning_rate_multiplier ⇒ Symbol, ...

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

Returns:

  • (Symbol, :auto, Float, nil) 


38
39
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 38

optional :learning_rate_multiplier,
union: -> { OpenAI::FineTuning::ReinforcementHyperparameters::LearningRateMultiplier }

#n_epochs ⇒ Symbol, ...

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

Returns:

  • (Symbol, :auto, Integer, nil) 


46
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 46

optional :n_epochs, union: -> { OpenAI::FineTuning::ReinforcementHyperparameters::NEpochs }

#reasoning_effort ⇒ Symbol, ...

Level of reasoning effort.

Returns:



52
53
54
55
 
# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 52

optional :reasoning_effort,
enum: -> {
  OpenAI::FineTuning::ReinforcementHyperparameters::ReasoningEffort
}

Class Method Details

.variants ⇒ Array(Symbol, :auto, Integer)

Returns:

  • (Array(Symbol, :auto, Integer)) 



# File 'lib/openai/models/fine_tuning/reinforcement_hyperparameters.rb', line 88