Interface DistillationSpec

Distillation tuning spec for tuning.

interface DistillationSpec {
    baseTeacherModel?: string;
    hyperParameters?: DistillationHyperParameters;
    pipelineRootDirectory?: string;
    promptDatasetUri?: string;
    studentModel?: string;
    trainingDatasetUri?: string;
    tunedTeacherModelSource?: string;
    validationDatasetUri?: string;
}

Properties

baseTeacherModel? hyperParameters? pipelineRootDirectory? promptDatasetUri? studentModel? trainingDatasetUri? tunedTeacherModelSource? validationDatasetUri?

Properties

OptionalbaseTeacherModel

baseTeacherModel?: string

The base teacher model that is being distilled. See Supported models.

OptionalhyperParameters

hyperParameters?: DistillationHyperParameters

Optional. Hyperparameters for Distillation.

OptionalpipelineRootDirectory

pipelineRootDirectory?: string

Deprecated. A path in a Cloud Storage bucket, which will be treated as the root output directory of the distillation pipeline. It is used by the system to generate the paths of output artifacts.

OptionalpromptDatasetUri

promptDatasetUri?: string

The GCS URI of the prompt dataset to use during distillation.

OptionalstudentModel

studentModel?: string

The student model that is being tuned, e.g., "google/gemma-2b-1.1-it". Deprecated. Use base_model instead.

OptionaltrainingDatasetUri

trainingDatasetUri?: string

Deprecated. Cloud Storage path to file containing training dataset for tuning. The dataset must be formatted as a JSONL file.

OptionaltunedTeacherModelSource

tunedTeacherModelSource?: string

The resource name of the Tuned teacher model. Format: projects/{project}/locations/{location}/models/{model}.

OptionalvalidationDatasetUri

validationDatasetUri?: string

Optional. Cloud Storage path to file containing validation dataset for tuning. The dataset must be formatted as a JSONL file.

Settings

On This Page

Properties