CreateTuningJobConfig

public abstract class CreateTuningJobConfig extends JsonSerializable

Fine-tuning job creation request - optional fields.

Constructors

Link copied to clipboard
public void CreateTuningJobConfig()

Types

Link copied to clipboard
public abstract class Builder
Builder for CreateTuningJobConfig.

Properties

Link copied to clipboard
public final static String MAX_READ_LENGTH_PROPERTY
System property to override the default max JSON string length (20MB) in read constraints.

Functions

Link copied to clipboard
public abstract Optional<AdapterSize> adapterSize()
Adapter size for tuning.
Link copied to clipboard
public abstract Optional<String> baseTeacherModel()
The base teacher model that is being distilled.
Link copied to clipboard
public abstract Optional<Integer> batchSize()
The batch size hyperparameter for tuning.
Link copied to clipboard
public abstract Optional<Float> beta()
Weight for KL Divergence regularization, Preference Optimization tuning only.
Link copied to clipboard
Instantiates a builder for CreateTuningJobConfig.
Link copied to clipboard
public abstract Optional<String> customBaseModel()
Custom base model for tuning.
Link copied to clipboard
public abstract Optional<String> description()
The description of the TuningJob
Link copied to clipboard
The encryption spec of the tuning job.
Link copied to clipboard
public abstract Optional<Integer> epochCount()
Number of complete passes the model makes over the entire training dataset during training.
Link copied to clipboard
Evaluation config for the tuning job.
Link copied to clipboard
If set to true, disable intermediate checkpoints and only the last checkpoint will be exported.
Link copied to clipboard
public static CreateTuningJobConfig fromJson(String jsonString)
Deserializes a JSON string to a CreateTuningJobConfig object.
Link copied to clipboard
@InternalApi()
public static T fromJsonNode<T extends JsonSerializable>(JsonNode jsonNode, Class<T> clazz)
Deserializes a JsonNode to an object of the given type.
Link copied to clipboard
@InternalApi()
public static T fromJsonString<T extends JsonSerializable>(String jsonString, Class<T> clazz)
Deserializes a Json string to an object of the given type.
Link copied to clipboard
public abstract Optional<HttpOptions> httpOptions()
Used to override HTTP request options.
Link copied to clipboard
public abstract Optional<Map<String, String>> labels()
Optional.
Link copied to clipboard
public abstract Optional<Float> learningRate()
The learning rate for tuning.
Link copied to clipboard
Multiplier for adjusting the default learning rate.
Link copied to clipboard
public abstract Optional<TuningMethod> method()
The method to use for tuning (SUPERVISED_FINE_TUNING or PREFERENCE_TUNING or DISTILLATION).
Link copied to clipboard
public static ObjectMapper objectMapper()
Link copied to clipboard
public abstract Optional<String> outputUri()
The Google Cloud Storage location where the tuning job outputs are written.
Link copied to clipboard
The optional checkpoint id of the pre-tuned model to use for tuning, if applicable.
Link copied to clipboard
public static void setMaxReadLength(int maxReadLength)
Overrides the default maximum JSON string length (20MB) for the JSON parser.
Link copied to clipboard
Multiplier for adjusting the weight of the SFT loss.
Link copied to clipboard
public static JsonNode stringToJsonNode(String string)
Converts a Json string to a JsonNode.
Link copied to clipboard
Creates a builder with the same values as this instance.
Link copied to clipboard
public String toJson()
Serializes the instance to a Json string.
Link copied to clipboard
public static JsonNode toJsonNode(Object object)
Serializes an object to a JsonNode.
Link copied to clipboard
public static String toJsonString(Object object)
Serializes an object to a Json string.
Link copied to clipboard
The display name of the tuned Model.
Link copied to clipboard
The resource name of the Tuned teacher model.
Link copied to clipboard
public abstract Optional<TuningMode> tuningMode()
Tuning mode for tuning.
Link copied to clipboard
Validation dataset for tuning.