PreferenceOptimizationDataStats

@Serializable
data class PreferenceOptimizationDataStats(val droppedExampleIndices: List<Long>? = null, val droppedExampleReasons: List<String>? = null, val scoreVariancePerExampleDistribution: DatasetDistribution? = null, val scoresDistribution: DatasetDistribution? = null, val totalBillableTokenCount: Long? = null, val tuningDatasetExampleCount: Long? = null, val tuningStepCount: Long? = null, val userDatasetExamples: List<GeminiPreferenceExample>? = null, val userInputTokenDistribution: DatasetDistribution? = null, val userOutputTokenDistribution: DatasetDistribution? = null)

Statistics computed for datasets used for preference optimization. This data type is not supported in Gemini API.

Constructors

Link copied to clipboard
constructor(droppedExampleIndices: List<Long>? = null, droppedExampleReasons: List<String>? = null, scoreVariancePerExampleDistribution: DatasetDistribution? = null, scoresDistribution: DatasetDistribution? = null, totalBillableTokenCount: Long? = null, tuningDatasetExampleCount: Long? = null, tuningStepCount: Long? = null, userDatasetExamples: List<GeminiPreferenceExample>? = null, userInputTokenDistribution: DatasetDistribution? = null, userOutputTokenDistribution: DatasetDistribution? = null)

Properties

Link copied to clipboard

Output only. A partial sample of the indices (starting from 1) of the dropped examples.

Link copied to clipboard

Output only. For each index in dropped_example_indices, the user-facing reason why the example was dropped.

Link copied to clipboard

Output only. Dataset distributions for scores.

Link copied to clipboard

Output only. Dataset distributions for scores variance per example.

Link copied to clipboard

Output only. Number of billable tokens in the tuning dataset.

Link copied to clipboard

Output only. Number of examples in the tuning dataset.

Link copied to clipboard
val tuningStepCount: Long? = null

Output only. Number of tuning steps for this Tuning Job.

Link copied to clipboard

Output only. Sample user examples in the training dataset.

Link copied to clipboard

Output only. Dataset distributions for the user input tokens.

Link copied to clipboard

Output only. Dataset distributions for the user output tokens.