feat: support hyperparameters in distillation tuning

PiperOrigin-RevId: 890622141

Author: speedstorm1

GeneratedFirebaseAI/Sources/Types.swift

@@ -20329,7 +20329,7 @@ extension PreferenceOptimizationSpec: Codable {
   }
 }
 
-/// Hyperparameters for Distillation. This data type is not supported in Gemini API.
+/// Hyperparameters for distillation.
 @available(iOS 15.0, macOS 13.0, macCatalyst 15.0, tvOS 15.0, watchOS 8.0, *)
 public struct DistillationHyperParameters: Sendable {
   /// Optional. Adapter size for distillation.
@@ -20342,15 +20342,26 @@ public struct DistillationHyperParameters: Sendable {
   /// Optional. Multiplier for adjusting the default learning rate.
   public let learningRateMultiplier: Double?
 
+  /// The batch size hyperparameter for tuning.
+  /// This is only supported for OSS models in Vertex.
+  public let batchSize: Int32?
+
+  /// The learning rate for tuning. OSS models only.
+  public let learningRate: Float?
+
   /// Default initializer.
   public init(
     adapterSize: AdapterSize? = nil,
     epochCount: Int64? = nil,
-    learningRateMultiplier: Double? = nil
+    learningRateMultiplier: Double? = nil,
+    batchSize: Int32? = nil,
+    learningRate: Float? = nil
   ) {
     self.adapterSize = adapterSize
     self.epochCount = epochCount
     self.learningRateMultiplier = learningRateMultiplier
+    self.batchSize = batchSize
+    self.learningRate = learningRate
   }
 }
 
@@ -20362,6 +20373,8 @@ extension DistillationHyperParameters: Codable {
     case adapterSize = "adapterSize"
     case epochCount = "epochCount"
     case learningRateMultiplier = "learningRateMultiplier"
+    case batchSize = "batchSize"
+    case learningRate = "learningRate"
   }
 
   public init(from decoder: any Decoder) throws {
@@ -20382,6 +20395,16 @@ extension DistillationHyperParameters: Codable {
       Double.self,
       forKey: .learningRateMultiplier
     )
+
+    batchSize = try VertexKeysContainer.decodeIfPresent(
+      Int32.self,
+      forKey: .batchSize
+    )
+
+    learningRate = try VertexKeysContainer.decodeIfPresent(
+      Float.self,
+      forKey: .learningRate
+    )
   }
 
   public func encode(to encoder: any Encoder) throws {
@@ -20405,6 +20428,16 @@ extension DistillationHyperParameters: Codable {
         forKey: .learningRateMultiplier
       )
 
+      try VertexKeysContainer.encodeIfPresent(
+        batchSize,
+        forKey: .batchSize
+      )
+
+      try VertexKeysContainer.encodeIfPresent(
+        learningRate,
+        forKey: .learningRate
+      )
+
     }
   }
 }
@@ -20444,6 +20477,9 @@ public struct DistillationSpec: Sendable {
   /// The dataset must be formatted as a JSONL file.
   public let validationDatasetUri: String?
 
+  /// Tuning mode for tuning.
+  public let tuningMode: TuningMode?
+
   /// Default initializer.
   public init(
     promptDatasetUri: String? = nil,
@@ -20453,7 +20489,8 @@ public struct DistillationSpec: Sendable {
     studentModel: String? = nil,
     trainingDatasetUri: String? = nil,
     tunedTeacherModelSource: String? = nil,
-    validationDatasetUri: String? = nil
+    validationDatasetUri: String? = nil,
+    tuningMode: TuningMode? = nil
   ) {
     self.promptDatasetUri = promptDatasetUri
     self.baseTeacherModel = baseTeacherModel
@@ -20463,6 +20500,7 @@ public struct DistillationSpec: Sendable {
     self.trainingDatasetUri = trainingDatasetUri
     self.tunedTeacherModelSource = tunedTeacherModelSource
     self.validationDatasetUri = validationDatasetUri
+    self.tuningMode = tuningMode
   }
 }
 
@@ -20479,6 +20517,7 @@ extension DistillationSpec: Codable {
     case trainingDatasetUri = "trainingDatasetUri"
     case tunedTeacherModelSource = "tunedTeacherModelSource"
     case validationDatasetUri = "validationDatasetUri"
+    case tuningMode = "tuningMode"
   }
 
   public init(from decoder: any Decoder) throws {
@@ -20524,6 +20563,11 @@ extension DistillationSpec: Codable {
       String.self,
       forKey: .validationDatasetUri
     )
+
+    tuningMode = try VertexKeysContainer.decodeIfPresent(
+      TuningMode.self,
+      forKey: .tuningMode
+    )
   }
 
   public func encode(to encoder: any Encoder) throws {
@@ -20572,6 +20616,11 @@ extension DistillationSpec: Codable {
         forKey: .validationDatasetUri
       )
 
+      try VertexKeysContainer.encodeIfPresent(
+        tuningMode,
+        forKey: .tuningMode
+      )
+
     }
   }
 }
@@ -24834,7 +24883,7 @@ public struct CreateTuningJobConfig: Sendable {
   /// Adapter size for tuning.
   public let adapterSize: AdapterSize?
 
-  /// Tuning mode for SFT tuning.
+  /// Tuning mode for tuning.
   public let tuningMode: TuningMode?
 
   /// Custom base model for tuning. This is only supported for OSS models in Vertex.