Package com.google.genai.types

Class ReinforcementTuningHyperParameters.Builder

java.lang.Object
com.google.genai.types.ReinforcementTuningHyperParameters.Builder
Enclosing class:
ReinforcementTuningHyperParameters
public abstract static class ReinforcementTuningHyperParameters.Builder extends Object
Builder for ReinforcementTuningHyperParameters.

  • Constructor Details

    • Builder

      public Builder()

  • Method Details

    • epochCount

      public abstract ReinforcementTuningHyperParameters.Builder epochCount(Long epochCount)
      Setter for epochCount.

      epochCount: Optional. Number of training epoches for the tuning job.

    • clearEpochCount

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearEpochCount()
      Clears the value of epochCount field.

    • learningRateMultiplier

      public abstract ReinforcementTuningHyperParameters.Builder learningRateMultiplier(Float learningRateMultiplier)
      Setter for learningRateMultiplier.

      learningRateMultiplier: Learning rate multiplier for Reinforcement Learning.

    • clearLearningRateMultiplier

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearLearningRateMultiplier()
      Clears the value of learningRateMultiplier field.

    • adapterSize

      public abstract ReinforcementTuningHyperParameters.Builder adapterSize(AdapterSize adapterSize)
      Setter for adapterSize.

      adapterSize: Optional. Adapter size for Reinforcement Tuning.

    • clearAdapterSize

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearAdapterSize()
      Clears the value of adapterSize field.

    • adapterSize

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder adapterSize(AdapterSize.Known knownType)
      Setter for adapterSize given a known enum.

      adapterSize: Optional. Adapter size for Reinforcement Tuning.

    • adapterSize

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder adapterSize(String adapterSize)
      Setter for adapterSize given a string.

      adapterSize: Optional. Adapter size for Reinforcement Tuning.

    • samplesPerPrompt

      public abstract ReinforcementTuningHyperParameters.Builder samplesPerPrompt(Integer samplesPerPrompt)
      Setter for samplesPerPrompt.

      samplesPerPrompt: Optional. Number of different responses to generate per prompt during tuning.

    • clearSamplesPerPrompt

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearSamplesPerPrompt()
      Clears the value of samplesPerPrompt field.

    • batchSize

      public abstract ReinforcementTuningHyperParameters.Builder batchSize(Integer batchSize)
      Setter for batchSize.

      batchSize: Optional. Batch size for the tuning job. How many prompts to process at a train step. If not set, the batch size will be determined automatically.

    • clearBatchSize

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearBatchSize()
      Clears the value of batchSize field.

    • evaluateInterval

      public abstract ReinforcementTuningHyperParameters.Builder evaluateInterval(Integer evaluateInterval)
      Setter for evaluateInterval.

      evaluateInterval: Optional. How often at steps to evaluate the tuning job during training. If not set, evel will be run per epoch. `total_steps = epoch_count * samples_per_prompt / total_prompts_in_dataset`

    • clearEvaluateInterval

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearEvaluateInterval()
      Clears the value of evaluateInterval field.

    • checkpointInterval

      public abstract ReinforcementTuningHyperParameters.Builder checkpointInterval(Integer checkpointInterval)
      Setter for checkpointInterval.

      checkpointInterval: Optional. How often at steps to save checkpoints during training. If not set, one checkpoint per epoch will be set. ```total_steps = epoch_count * samples_per_prompt / total_prompts_in_dataset```

    • clearCheckpointInterval

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearCheckpointInterval()
      Clears the value of checkpointInterval field.

    • maxOutputTokens

      public abstract ReinforcementTuningHyperParameters.Builder maxOutputTokens(Integer maxOutputTokens)
      Setter for maxOutputTokens.

      maxOutputTokens: Optional. The maximum number of tokens to generate per prompt. Default to 32768.

    • clearMaxOutputTokens

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearMaxOutputTokens()
      Clears the value of maxOutputTokens field.

    • thinkingLevel

      public abstract ReinforcementTuningHyperParameters.Builder thinkingLevel(ReinforcementTuningThinkingLevel thinkingLevel)
      Setter for thinkingLevel.

      thinkingLevel: Indicates the maximum thinking depth during tuning. Starting from Gemini 3.5 models, the old thinking_budget will no longer be supported and will result in a user error if set. Instead, users should use the thinking_level parameter to control the maximum thinking depth.

    • clearThinkingLevel

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearThinkingLevel()
      Clears the value of thinkingLevel field.

    • thinkingLevel

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder thinkingLevel(ReinforcementTuningThinkingLevel.Known knownType)
      Setter for thinkingLevel given a known enum.

      thinkingLevel: Indicates the maximum thinking depth during tuning. Starting from Gemini 3.5 models, the old thinking_budget will no longer be supported and will result in a user error if set. Instead, users should use the thinking_level parameter to control the maximum thinking depth.

    • thinkingLevel

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder thinkingLevel(String thinkingLevel)
      Setter for thinkingLevel given a string.

      thinkingLevel: Indicates the maximum thinking depth during tuning. Starting from Gemini 3.5 models, the old thinking_budget will no longer be supported and will result in a user error if set. Instead, users should use the thinking_level parameter to control the maximum thinking depth.

    • thinkingBudget

      public abstract ReinforcementTuningHyperParameters.Builder thinkingBudget(Integer thinkingBudget)
      Setter for thinkingBudget.

      thinkingBudget: Optional. The thinking budget for the tuning job to optimize for (Gemini 2.5 only). * -1 means dynamic thinking * 0 means no thinking * > 0 means thinking budget in tokens If not set, default to -1 (dynamic thinking).

    • clearThinkingBudget

      @CanIgnoreReturnValue public ReinforcementTuningHyperParameters.Builder clearThinkingBudget()
      Clears the value of thinkingBudget field.

    • build

      public abstract ReinforcementTuningHyperParameters build()